MODELING ATRIAL FIBRILLATION USING
HUMAN EMBRYONIC STEM CELLS
Zachary Laksman
A thesis submitted in conformity with the requirements for the
Degree of Masters of Science
Institute of Medical Sciences
University of Toronto
© Copyright by « Zachary Laksman » « 2015»
ii
Abstract
Thesis Title: Modeling Atrial Fibrillation Using Human Embryonic Stem Cells
Degree: Masters of Science, Institute of Medical Sciences, University of Toronto
Convocation: November 2015
Background: Atrial fibrillation (AF) is the most common clinical arrhythmia and is
associated with significant morbidity and mortality.
Hypothesis: “Atrial” cardiomyocytes generated from hESCs will recapitulate the key
hallmarks of human AF in a 2 dimensional tissue model, as well as demonstrate the
expected electrophysiologic changes in response to drugs.
Methods and Results: Our “atrial” cardiomyocytes were 80-90% positive for cardiac
troponin, were enriched for the atrial markers ANF and KCNJ3, and lacked ventricular
markers (MYL2 and IRX4). The “atrial” cells mimicked the human electrophysiologic
phenotype with the majority (90%) demonstrating atrial like action potentials. Optical
mapping of multicellular cell sheets was performed on rotors at baseline, after induction
of rotor formation and with the addition of commonly employed antiarrhythmic drugs.
Conclusions: We have successfully generated a human model of AF in vitro and
demonstrated its utility in screening for known, and unpredicted effects of commonly
employed anti-arrhythmics.
iii
Acknowledgements
I would like to thank Dr. Peter Backx and Dr. Gordon Keller for their support and mentorship. It has been a great privilege to work with scientists of the highest caliber, and people who have earned my greatest esteem and respect. I would like to thank my beautiful daughters Kaiya, Nava, Arielle, and Samara. Thank you to all my family and friends for their help and understanding. Finally, thanks to my wife Jodi who remains the strongest person I have ever met. Thank you for continuing to believe in me and support me unconditionally. I would like to dedicate this thesis to my late father Sander Laksman.
iv
Technical Contributions and Acknowledgement I acknowledge the following individuals for their contributions to my MSc project: Stephanie Protze, PhD. and Jeehoon Lee, PhD candidate Dr. Protze and Jeehoon’s work in the development and validation of the differentiation protocol used to derive the “atrial” cells under study laid the foundation upon which this project was based. They performed the qRT-PCR and aldefluor stains reported in this thesis as a part of these efforts. Wallace Yang Worked tirelessly to optimize and improve the optical mapping rig and its requisite signal processing. Dr. Mark Gagliardi, PhD Provided assistance and expert consultation in all aspects of tissue culture and directed differentiation of hPSCs Dr. Roozbeh Aschar-Sobbi, PhD Provided expert consultation on nearly all aspects of experimental cardiac electrophysiology Farzad Izaddoustdar Assisted in signal processing and technical aspects related to optical mapping
v
Table of Contents
Abstract
ii
Acknowledgements iii
Technical contributions and acknowledgements iv
List of abbreviations vii
List of Tables viii
List of Figures ix
Chapter 1 Background 1 1.1 Atrial fibrillation (AF) – clinical burden 1 1.2 Normal cardiac physiology 1 1.3 AF management 3 1.3.1 Rate control vs. rhythm control of AF 3 1.3.2 AF ablation 5 1.4 Basic electrical properties and mechanisms underlying AF 7 1.5 Anti-arrhythmic drugs: dofetilide and flecainide 11 1.6 Wave theory 12 1.7 Rotors 13 1.7.1 Physiology of rotors 14 1.7.2 Clinical relevance of rotors in AF 16 1.8 Embryonic stem cells and induced pluripotent stem cells 18 1.8.1 Directed Differentiation 20 1.8.2 Retinoic acid signaling 23 1.8.3 Molecular markers of atrial and ventricular cardiomyocytes 24 1.9 Disease modeling using hPSCs 25 1.9.1cLQTS 26 1.9.2 acquired LQTS 29
Chapter 2 Research Aims and Hypotheses 33
Chapter 3 Materials and Methods 35 3.1 Differentiation protocols 35 3.2 Flow cytometry and cell sorting 36 3.3 Molecular markers 37 3.31 qRT-PCR 37 3.3.2 Immunostaining 37 3.4 Cardiac Electrophysiology 37 3.4.1 Patch clamping 37 3.4.2 Micro Electrode Array 39 3.4.3 Optical mapping 40 3.4.3.1 Signal processing 42 3.4.4 Intracellular recordings of cell sheets 45
vi
Chapter 4 Results
4.1 Differentiation protocols of “atrial” and “ventricular” cardiomyocytes 46 4.2 Molecular markers of “atrial” and “ventricular” cells 51 4.3 Single cell electrophysiology studies of “atrial” and “ventricular” cells 53 4.4 MEA studies of “atrial” and “ventricular” cell sheets 57 4.5 Optical mapping studies of “atrial” and “ventricular” cell sheets 59 4.5.1 Optical mapping the effects of antiarrhythmic drugs: flecainide and dofetilide
65
Chapter 5 Discussion 72 5.1 Molecular markers of “atrial” and “ventricular” cardiomyocytes 73 5.2 Single cell electrophysiology 73 5.3 Modeling atrial disease using hPSCs 75 5.3.1 APD restitution and heterogeneity 78 5.3.2 Conduction velocity 79 5.3.3 Effects of dofetilide on “atrial” cell sheets 80 5.3.4 Effects of flecainide on “atrial” cell sheets 81
Chapter 6 Future Directions 87 6.1 Validation of current findings 87 6.2 Study AF remodeling 90 6.3 Model other types and features of acquired AF 92 6.4 Introduce genetic variation as a determinant and model of AF 94
Chapter 7 References 99
Chapter 8 Appendix 122
vii
List of abbreviations
AAD Anti-arrhythmic drug
AF Atrial fibrillation
ANF Atrial natriuretic factor
AP Action potential
APA Action potential amplitude
APD Action potential duration
APD50 Action potential duration at 50% of repolarization
APD90 Action potential duration at 90% of repolarization
CaV1.3 Alpha 1D subunit of the L-type voltage dependent calcium channel
CM Cardiomyocyte
CV Conduction velocity
Cx Connexin
DMP Diastolic membrane potential
dv/dtmax Maximum action potential upstroke velocity
EAD Early after depolarizations
EB Embryoid body
ECG Electrocardiogram
ERP Effective refractory period
ESC Embryonic stem cell
FDA U.S. Food and Drug Administration
FFT Fast fourier transform
GJ Gap junction
hPSC Human pluripotent stem cell
ICD Implantable cardioverter defibrillator
iPSC Induced pluripotent stem cell
KCNJ3 Potassium inward rectifying channel, subfamily J, member 3
LQTS Long QT syndrome
MEA Micro electrode array
MYL2 Myosin regulatory light chain-2
NCX Sodium Calcium exchanger
OAPD Optical action potential duration
PS Phase singularity
PV Pulmonary vein
qRT-PCR Real-time quantitative reverse transcription polymerase chain reaction
RA Retinoic acid
RMP Resting membrane potential
ROI Region of interest
RP Refractory period
SAN Sinoatrial node
SR Sinus rhythm
TDR Transmural dispersion of repolarization
TdP Torsades-de-pointes
viii
List of Tables
Table 1: Characteristics of APs recorded from the protocol generating “atrial” cardiomyocytes compared to the protocol used to generate “ventricular” cardiomyocytes.
55
ix
List of Figures
Figure 1: Flow cytometric analyses of markers of undifferentiated ES cell 46 Figure 2: Scheme of the protocol used to differentiate hESCs into “atrial” and “ventricular” cardiomyocytes
47
Figure 3: qRT-PCR results comparing relative expression of ALDH1A2 and
CYP26A1 in low induction (BMP4 3 ng/ml and Activin A 2 ng/ml) compared
to high induction (BMP4 10 ng/ml and Activin A 6 ng/ml)
48
Figure 4: Flow cytometric analyses of low induction vs. high induction with
respect to PDGFRα and Aldefluor staining
49
Figure 5: Flow cytometric analyses of important markers used in the
optimization and validation of the directed differentiation protocol (CD56,
PdgfR-α, CD90, SIRPα, cTNT).
51
Figure 6: qRT-PCR-based expression analyses of markers of atrial and
ventricular cardiomyocytes in “atrial” and “ventricular” differentiation
protocols
52
Figure 7: Fluorescent immunostaining of cTNT and the ventricular specific
marker MLC2v in “atrial” and “ventricular” differentiation protocols
53
Figure 8: Typical patch recordings of APs generated from “atrial” and
“ventricular” differentiation protocols
54
Figure 9: Showing presence of cardiomyocyte types with representative APs
and their prevalence in the “atrial” differentiation protocol
54
Figure 10: Showing presence of cardiomyocyte types with representative
APs and their prevalence in the “ventricular” differentiation protocol
55
Figure 11: Comparing the AP characteristics of cells generated from the “atrial” differentiation protocol compared to the “ventricular” differentiation protocol.
56
Figure 12: Representative patch clamp recordings of an anode break in
“atrial” cardiomyocytes
57
Figure 13: Representative MEA recordings of FPDs from “atrial” and “ventricular” cell sheets
58
Figure 14: MEA recording demonstrating pacing and capture of a cell sheet 58 Figure 15: Time sequence imaging of an “atrial” cell sheet in “SR” 60
Figure 16: Representative example of typical OAPs generated from an “atrial” cell sheet
60
Figure 17: On the left, an APD restitution curve of an “atrial” cell sheet
generated by plotting OAPD vs. rate. On the right, a recording of electrical
alternans generated in an “atrial” cell sheet during burst pacing.
61
Figure 18: Time sequence imaging of an “atrial” cell sheet in “AF” 61 Figure 19: APD maps of “SR” and “AF” of “atrial” cell sheets 62 Figure 20: Representative activation maps of “SR” and “AF” of “atrial” cell
sheets 64
Figure 21: Representative conduction velocity maps of “SR” and “AF” of “atrial” cell sheets
64
Figure 22: Representative examples of OAPDs and the effects of dofetilide and flecainide on OAP morphology in “atrial” cell sheet
65
x
Figure 23: Effect of rotor induction and dofetilide on cycle length and APD in “atrial” cell sheets
66
Figure 24: Effect of rotor induction and dofetilide on conduction velocity in
“atrial” cell sheets
67
Figure 25: Effect of flecainide on cycle length and APD in “atrial” cell sheets 68 Figure 26: Effect of flecainide on conduction velocity in “atrial” cell sheets 69
Figure 27: Effect of rotor induction and dofetilide on conduction velocity in
“atrial” cell sheets as a function of the distance from the origin of the
electrical wavefront
70
Figure 28: Effect of flecainide on conduction velocity in “atrial” cell sheets as
a function of the distance from the origin of the electrical wavefront
71
Appendix Figure 1: Example of an intracellular recording of a cell sheet using
high resistance micropipettes
122
Appendix Figure 2: Serial activation maps generated after the induction of a
rotor in an “atrial” cell sheet demonstrating the initiation of two rotors after
the application of flecainide
123
Appendix Figure 3: Optical APs recorded from hiPSC derived
cardiomyocytes expressing Arclight.
124
1
Chapter 1
Background
1.1 Atrial Fibrillation – clinical burden
Atrial fibrillation (AF) is the most common clinical arrhythmia, currently affecting 1-2%
of the total population 1. AF is the most common arrhythmia necessitating hospital
admission, and is a major and growing burden on the health of Canadians and our health
care system. Age is an important determinant of AF occurrence with a 25% lifetime risk
of developing AF after the age of 40. As the proportion of our populations get older,
along with a rise in a set of parallel AF risk factors, the prevalence of AF is expected to
double over the next 4 decades2. A diagnosis of AF is associated with a doubling in all-
cause mortality, a five-fold increase in stroke, an accelerated development of heart
failure, and a substantially poorer quality of life3,4. In addition to the costs of identifying
and treating cardiac risk factors leading to AF5-8, costs directly attributed to AF exceeded
$20,000/year/patient in 20109. Total costs are expected to accelerate with more
widespread use of expensive, and invasive, AF ablation ($20k/procedure)10, which can be
seen as palliation rather than a cure, touting success rates of less than 60% after 1 year
even in highly selected patient populations11.
1.2 Normal Cardiac Physiology
To begin to discuss the pathophysiology that predisposes patients to AF, we must first
review normal cardiac physiology, and the link between its conduction system and force
generating contraction. While all myocytes within the heart have the capacity to conduct
electrical impulses through gap junctions, the heart’s pump function under normal
conditions is tightly regulated by a specialized conduction system. When an electrical
2
impulse (action potential) stimulates myocardium it contracts. This process, which is
fundamental to all muscle physiology, is known as excitation-contraction coupling. The
conduction system of the atria and ventricles are designed to undergo regular and orderly
depolarization/repolarization sequences in order to maximize cardiac efficiency while
minimizing arrhythmias. The sinoatrial node (SAN) acts as the primary pacemaker of the
heart, depolarizing spontaneously to determine and drive the rate of cardiac contraction.
This normal physiologic rhythm is called sinus rhythm (SR). The electrical wavefront
generated by the SAN first initiates the coordinated contraction of the right and left atria.
Atrial contraction ejects blood in the ventricles, the bottom chambers of the heart,
through the atrioventricular valves, contributing as much as 30% towards the final cardiac
output ultimately ejected by the ventricles through the semilunar valves. After atrial
depolarization, the electrical wavefront then invades the atrioventricular node, which
serves as a gateway to the specialized conduction system of the ventricles called the His-
Purkinje system. When AF is present, atria no longer beat regularly (i.e. atria "fibrillate").
AF is easily detected on electrocardiographic recordings (ECGs) as rapid, "irregularly-
irregular" QRS complexes (from ventricles) along with irregular (“random”) baseline
signals (from atria)12,13.
Within the atria, AF is characterized by the appearance of abnormal electrical re-entry
circuits which are self-sustained sequences of depolarization-repolarization cycles called
rotors or wavelets14-16. Despite cyclic patterns seen with re-entry, the electrical patterns in
AF usually have “random” signatures due to wave-breakup/conduction blocks, thus the
term "wavelets" has been proposed14. Crudely speaking, though important for AF
discussion, the propensity towards re-entry (rotors/wavelets) depends on an electrical
3
property called the “atrial wavelength" (i.e. λ) which is the product of the effective
refractory period (ERP), which is ~action potential duration (APD), and the conduction
velocity (CV), which is the speed of spreading action potentials (APs) in atria. Re-entry
however does not occur without initiation by processes called “kindling” or "triggering".
Both the substrate of AF, the atrial wavelength, and the triggers of AF are potential
targets for therapy.
1.3 AF management
Although atrial beating normally makes relatively minor contributions to the heart's
overall pumping efficiency (approximately 15%), AF nevertheless has important clinical
consequences. For one, because ventricles are electrically-entrained by the atria via the
AV-node, AF leads to elevated pumping rates of the heart (typically >120 beats/min
when AF is untreated)14 which severely impairs cardiac output regulation, leading to
patient fatigue, and promoting “tachycardia induced” cardiomyopathy. The loss of
sequential atrial contractions also causes blood pooling within atrial appendages resulting
in thrombus formation, embolization and stroke, which is the major cause of morbidity
and mortality in AF patients17,18. Antithrombotic therapy is thus recommended for a
majority of patients. The decision regarding initiation of anticoagulation balances the
patient specific risks of bleeding against the treatment and patient specific benefits of
antiplatelet and/or anticoagulants 19. This decision is made independent of the choice
regarding the management of the underlying arrhythmia.
1.3.1 Rate control vs. rhythm control
There are two main strategies for the electrical management of AF, rhythm control and
rate control. Despite the intuitive attraction of restoring sinus rhythm, several trials have
4
failed to show clinical benefit of this strategy over rate control. The rate control strategy
involves the use of AV nodal blocking agents, such as beta blockers, calcium channel
blockers, and digoxin, to slow impulses through the AV node while allowing
uninterrupted fibrillation of the atrial chambers. Several studies have shown non-
inferiority of rhythm control compared to rate control in terms of survival benefit 20-23,
while studies targeting a higher risk population of heart failure patients similarly did not
show clinical benefit to the rhythm control strategy 24,25. Because of this clinical trial
data, rhythm control is only indicated for patients who suffer from symptoms of AF, or
those who cannot receive adequate rate control and suffer from tachycardia induced
cardiomyopathy 19.
It is worth dissecting some of the trial data that led to our current guideline
recommendations further when considering avenues for future research and novel
therapeutics. Prior to the AFFIRM trial in 2002, a study which arguably has had the most
significant impact on the clinical management of AF, first line therapy was rhythm
control with an implicit belief that patients would benefit from fewer symptoms, better
exercise tolerance, a lower risk of stroke, eventual discontinuation of long-term
anticoagulant therapy and better survival21. In this trial of over 4000 patients over the age
of 65, and including 25% of patients with decreased left ventricular systolic function, the
rhythm control strategy was left to the discretion of the primary physician and included
anti-arrhythmic drugs (AADs) associated with increased mortality, particularly in this
type of high risk population group. The study determined that there was a trend towards
worse survival and increased hospitalizations in patients on a rhythm control compared to
a rate control strategy. Further complicating the picture, patients on rhythm control had a
5
higher propensity towards discontinuing their anti-coagulation. Interestingly, 37.4% of
patients in the rhythm control group remained in AF, and 34.6% of patients in the rate
control group were in sinus rhythm, clearly demonstrating the ineffectual nature of the
era’s anti-arrhythmic strategy, and supporting the concept that this was not truly a
comparison of successful rhythm control compared to rate control. Furthermore,
crossover between groups was moderately large with 15% of rate control patients and
38% of rhythm control patients crossing over to the other treatment group 21. Later, an “
on-treatment” analysis of the AFFIRM trial showed that the presence of sinus rhythm was
in fact associated with a lower risk of death 26.
Further post-hoc analyses of rhythm vs. rate control strategies have shown a consistent
pattern of improvement in quality of life measures in patients who achieve and maintain
sinus rhythm 27,28. In the PIAF and HOT CAFE trials for example, 6-minute walk times
and maximal treadmill workloads were increased 20,29. In the Canadian Trial of Atrial
Fibrillation, quality of life measures were significantly improved at 3 months from
baseline in patients who achieved sinus rhythm 30.
1.3.2 AF ablation
From the aforementioned results, two main streams of thought have become pervasive in
the field. The first continues to target symptoms with little emphasis on the
electrocardiographic documentation of sinus rhythm or AF. The second centers on a
belief that to date, there has never been a true comparison of rate vs. rhythm control on
“hard” outcomes such as hospitalization, progression of left ventricular dysfunction, or
mortality, since we have not had a therapy that can successfully maintain a high
proportion of patients in sinus rhythm to act as a reasonable on-treatment group. One of
6
the hopes for a more effective and durable therapy has been the invasive management of
catheter ablation. Remarkably, the first catheter ablation of AF was performed in 1981 by
Dr. Scheinman using high energy DC shocks to destroy the AV node as a form of rate
control 31. Since that time, the technology has rapidly evolved to include radiofrequency
energy that heats the tip of the catheter to a much higher temperature and in a
considerably more controlled fashion. The most significant advancement was
subsequently made in 1998 when Dr. Haissaguerre found that pulmonary vein (PV)
ectopy, originating from the muscular sleeves that extend from the atrial myocardium, are
frequent triggers for AF that can be successfully targeted for ablation 32. They found that
94% of the ectopic foci responsible for initiating AF originated from the PV in a group of
45 patients. Long term success rates from this procedure were 33% off AADs with an
additional 13% of patients being free of AF on AADs33. In 2001, it was Dr. Haissaguerre
again who described a novel approach to PV isolation that targeted the proximal insertion
points of the PVs to the left atrium, thus decreasing the most notable complication of PV
ectopy (PV stenosis), and providing benefit to patients who did not have inducible PV
ectopy during their electrophysiology study.
Since 2001, incremental improvements in catheter and mapping technology have
certainly improved safety and efficacy of the procedure, however the success rates remain
suboptimal. The procedure predominantly targets the triggers of AF, and to date, there
has been little convincing evidence that ablation can successfully target the underlying
substrate34. Therefore, the best results from ablation are in those with paroxysmal AF of
short duration, absence of cardiac structural disease and comorbidities, and without signs
of atrial remodeling, such as dilatation or fibrosis, as seen on MRI using late gadolinium
7
enhancement 35-41. A meta-analysis of 19 studies that targeted this relatively small
population demonstrated a 68% success rate at 1 year and 60% success at maintaining
sinus rhythm at 3 years 11.
Despite the steady improvements in catheter ablation techniques, they remain expensive,
high risk procedures, which target only patients with lone paroxysmal AF, a relatively
small subpopulation of the growing population of patients with AF. Furthermore, while
AF invariably begins as isolated short-lived episodes called "paroxysmal AF", each
paroxysm accelerates atrial remodelling thereby promoting "persistent AF" (i.e. "AF-
begets-AF")42,43. As discussed previously, persistent AF is much less amenable to the
current ablation strategies, with abysmal outcomes and a lack of consensus regarding the
ability of any ablation strategy to modify the underlying substrate. For these reasons, and
the fact that there is growing evidence that maintenance of sinus rhythm is an important
consideration of outcome studies, there is renewed interest in the development of AADs
for the treatment of AF. An effective drug would not only be applicable to a wider
population, but may benefit the population of patients waiting up to 1 year for their AF
ablation in decreasing AF related remodelling which has been shown to predict worse
outcomes with ablation.
1.4 Basic electrical properties and mechanisms underlying AF
To discuss the future of AAD therapy for AF, we must first revisit the concepts alluded to
earlier, namely the basic mechanisms underlying atrial re-entry. As already mentioned,
AF generally is believed to require a vulnerable substrate that generally requires a
kindling process to initiate semi-random electrical activity featuring underlying "re-entry"
events. The fundamental appreciation of the functional determinants of AF are critical
8
for understanding the mechanism of action of current AADs. Current dogma proposes
that the generation (triggers) and maintenance of continuous or fibrillatory activity in the
atria depends on the underlying properties of the atrial (and related regions like the
pulmonary veins) substrate and the balance between the determinants of refractoriness
and excitability, or the imbalance thereof. As discussed earlier, the propensity towards
re-entry in the atria depends on the atrial wavelength (i.e. λ) which is the product of the
effective refractory period (ERP), which is ~action potential duration (APD), and the
conduction velocity (CV), which is the speed of spreading APs in atria (λ = RP X CV). If
λ is small relative to the physical dimensions of atria44 then re-entry circuits are more
likely, and hypothetically the atria could accommodate a larger number of simultaneous
reentry circuits.
The action potential, a key component of the atrial wavelength, constitutes changes in the
membrane potential of cardiomyocytes. The membrane potential is established by an
unequal distribution of electrically charged ions across the sarcolemma and
predominantly by the presence of conducting ion channels in the sarcolemma. Opening
and closing of the ion channels allows ionic currents to flow across the membrane
culminating in the generation of the action potential. The direction of current depends on
the electrochemical gradient of the corresponding ions, and the conductivity of channels
that carry the current alters at different membrane potentials.
The atrial action potential is comprised of 5 phases. Depolarization (phase 0), a short-
lived hyperpolarization (phase 1), the plateau (phase 2), and repolarization (phase 3) to
resting potentials (phase 4). Activation (phase 0) is primarily driven by the depolarizing
inward Na+ current (INa). Upon reaching threshold, the rapid influx of Na+ through
9
voltage-gated sodium channels causes further rapid depolarization of the membrane
(typically to +30 to +40mV), which in turn activates the transient outward potassium
current (Ito), causing very rapid, but short-lived repolarization of the membrane potential
(phase 1). Activation of L-type Ca2+ current (ICaL) follows with simultaneous activation
of a several repolarizing voltage-gated K+ channels (IK). The plateau phase (phase 2) of
the action potential and therefore the action potential duration (APD) is governed by a
delicate balance of ICaL and IK, with membrane repolarization being driven ultimately by
inactivation of L-type Ca2+ channels and progressive activation of IK (phase 3). The
resting potential (typically at -70 to -80 mV) is maintained by the inward rectifier current
IK1 and Ca2+ extrusion is carried out primarily by the electrogenic Na2+/Ca2+
exchanger (NCX) current (INCX). NCX exchanges 3Na2+ ions for each Ca2+ ion with one
net positive charge moving in the direction of sodium transport. Ion channels that are
enriched in atrial cardiomyocytes compared to other types of cardiomyocytes contribute
to the atrial specific action potential morphology. In contrast to atrial and ventricular
myocytes, SAN and AVN myocytes demonstrate slow depolarization of the resting
potential during phase 4, a property related to the relative absence of Ik1 allowing inward
currents (predominantly If) to depolarize the membrane potential 45. Slow depolarization
during phase 4 inactivates most sodium channels and thus depolarization is mainly
achieved by IcaL and T-type Ca2+ currents. INa reductions46-49 and increases in several K+
currents are linked to AF44,46,50-54. APD heterogeneity55-58 can be altered by either pacing
or vagal stimulation leading to regional differences in muscarinic-activated K+ currents,
IK,Ach16,57,59-61. It is important to note that changes in resting membrane potentials, which
10
are determined by "background" K+ currents (i.e. IK1 and IK,Ach) can also contribute to AF
by modulating INa inactivation and availability.
Conduction Velocity is a measure of the spread of depolarization between electrically
coupled regions within the atria and is determined primarily by sodium current (INa)
densities/kinetics and gap junction (GJ) channel properties62-64. In atria, INa is largely
generated by Nav1.5 pore-forming α-subunits, coassembled with various β-subunits
(Nav) and other proteins65 . GJs are clusters of closely packed channels that directly
connect the cytoplasmic compartments of adjacent cells allowing the passage of ions and
small molecules 66. Each of the neighbouring cardiomyocytes contributes 1 connexon, of
which 21 members have been identified. In the atria, GJ channels are formed by both
Connexin40 (Cx-40) and Connexin-43 (Cx-43) proteins which co-localize to a
considerable extent 62,67. Cx-40 is specific for the atria, and important for atrial
conduction. Targeted deletion of Cx-40 in mice generates a phenotype of diminished
atrial conduction velocity (up to 30%) 68. Somatic mutations in the Cx-40 gene (GJA5) in
humans is associated with a heritable form of AF, thought to be related to reduced
intercellular electrical coupling 69. Taken together, changes in the expression (reductions),
distribution and phosphorylation (reductions) of GJ proteins and INa are associated with
AF46,48,49,70-74. Note the contributors to RP and CV are chamber specific and reflect the
concert of ion channels and gap junctions specific to the chamber of interest.
Since shorter wavelengths promote AF induction and maintenance, many anti-
arrhythmics target RP and thus wavelength prolongation in order to suppress AF.
Alternatively, and less important for this discussion, AADs can act by suppressing AF
11
triggers. For the purpose of this study, we will discuss two anti-arrhythmics, dofetilide
and flecainide, in further depth.
1.5 Anti-arrhythmic drugs: dofetilide and flecainide
Dofetilide is a class III anti-arrhythmic which selectively inhibits the rapid component of
the time-dependent outward potassium current (Ikr) 75. Studies in dogs demonstrated
prolongation of the ERP and APD in a dose dependent manner, as well as dofetilide’s
ability to facilitate conversion of electrically induced fibrillation 76,77. Dofetilide has no
effect on the maximum rate of depolarization and does not influence conduction within
the His-Purkinje system or within the myocardium75. Data from placebo controlled trials
demonstrated the efficacy of dofetilide 78 with a 30% conversion rate to sinus rhythm, and
a 60% success rate at maintaining sinus rhythm out to 1 year, compared to 20% in the
placebo group. In patients with structural heart disease, dofetilide had no effect on
mortality, and decreased the risk of hospitalization and progression of congestive heart
failure 79,80. As predicted however, since Ikr is an important contributor to the
repolarization of ventricular cells, many patients were exposed to the pro-arrhythmic
nature of ventricular APD prolongation, which can lead to QT prolongation and put
patients at risk for torsades-de –pointes. In clinical trials, the incidence of torsades was as
high as 5% leading to a significant decrease in the clinical uptake of the drug, as well as
explicit warnings and regulations regarding its use 81,82.
Flecainide is a class Ic antiarrhythmic agent, classically considered to slow CV with little
effect on ERP 83,84This in itself would not be predicted to have favourable effects on AF
as it would decrease wavelength and increase AF susceptibility. Flecainide however has
been shown in dog models to increase the atrial refractory period in a rate dependent
12
fashion thus increasing wavelength at high rates 58. This has been attributed to the drug’s
effect on decreasing atrial APD accommodation during periods of increased heart rate,
possibly related to an effect on repolarizing potassium current, as well as decreasing the
heterogeneity of atrial activation 85. Flecainide has been shown in placebo controlled
trials to decrease AF recurrence once converted to sinus rhythm 86,87. Flecainide has
known pro-arrhythmic properties and has been associated with increased mortality in
patients with a history of ischemic heart disease, and left ventricular dysfunction 88-90.
Increased inducibility of ventricular arrhythmias in animal models of infarction and
ventricular dysfunction have been reproducibly observed 91. This has been attributed in
part to paradoxical amplification of flecainide induced conduction slowing in depolarized
tissue producing dispersion of conduction 91.
1.6 Wave theory
To this point, the discussion regarding AF substrate has ignored a critical concept in
arrhythmia induction and perpetuation by presuming that there is an anatomic obstacle at
the center of our reentrant circuit. To fill in this gap in our developing theorem of AF,
once again historical context will shed light on the evolving concepts. In fact, studies of
fibrillation and its mechanisms date back to the turn of the 20th century when Mayer
demonstrated sustainable circulatory activity in isolated rings of the contractile bell of the
rhyzostomous Scyphomedusa (jellyfish) and turtle ventricular muscle 92. By applying
electrical pulses at one end of the ring, Mayer induced activation wavefronts that
circulated uni-directionally around the ring. This was reproduced by Mines and Garrey
and then described as the circus movement hypothesis of reentry by Lewis in which he
described a reentrant arrhythmia as a wavefront that circulated back to its partially
13
refractory tail 93 94. Wave theory continued to develop in the 1960s and 1970s, driven
predominantly by technological innovations such as high-speed cameras and the
application of computer or numerical modeling. In 1973 Alessie experimentally modeled
reentry in the absence of an anatomical obstruction thus introducing the concept of
reentry around a functional obstacle or a rotor 95. Using microelectrode recordings in an
isolated dog atrial preparation, they demonstrated sustained reentry in healthy muscle.
This led to the leading circle model of functional reentry which relied on a center of
refractoriness around which the rotating wave circulates, maintained in its state of
refractoriness through ongoing bombardment of centripetal wave fronts. In contrast to
anatomical reentry, the leading circle hypothesis proposes that there is no fully excitable
gap as the circulating wavefront must encroach on its own tail.
1.7 Rotors
Interestingly, circulating wavefronts are not unique to cardiac tissue. The concept of
rotors underpinning cardiac fibrillation stems from the field of nonequilibrium
thermodynamics. The discovery of the initial phenomenon birthing this field is attributed
to Boris Belousov who observed in the 1950s oscillations of colours in a colourless
solution of potassium bromate, cerium sulfate, malonic acid and citric acid in sulfuric
acid 96. Later, Anatoly Zhabotinsky rediscovered this reaction sequence and ultimately
disseminated the concept widely at a conference in 1968 97. The oscillatory patterns
observed in the “BZ” reaction are seen across many natural phenomena including the
growth pattern of amoeba.
14
1.7.1 Physiology of rotors
A rotor is similar to the leading edge model of functional reentry with a critical
difference, the curved wavefront must meet with its wavetail at a singularity where the
tissue is not refractory 98. Rotors are hypothesized to be the drivers of cardiac fibrillation
and can be represented in 2-dimensional (2D) space as spiral waves or in 3-D space as
scroll waves 92. The wavefront represents cells that have undergone full excitation
(depolarization) and are returning to their rest state (repolarization). Gray et al
demonstrated a phase singularity (PS) using optical mapping experiments of fibrillation
in 1998, and generated phase maps which could track the spiral, its curvature, and its tip
over time and space 99. This was a dramatic step forward in the study of rotor dynamics
allowing its visualization and quantification. They demonstrated that the PS anchored the
rotor, while the spiral wave rotated around it. A rotor however, unlike the leading circle
model, allows for changes in the location of the PS dependent on subtle changes in the
rotor’s environment, or chaotic unpredictable behavior generating complex shapes.
Principal to the rotor theory is the curvature of the rotating wavefront which controls the
velocity of the impulse and the dynamics of the reentrant wavefront 100.
The leading front is curved, in contrast to the leading circle model of reentry, which gives
the activation a spiral shape. An important concept to consider is the source/sink
dynamics of a propagating wavefront. For excitation to occur, a depolarizing wave (the
source) must carry sufficient current to bring membrane potentials to threshold for firing
(activating voltage-gated sodium channels) in electrically vulnerable “downstream cells”,
which act as a sink 101 102. The geometry of the wavefront influences the source/sink ratio.
Compared to a planar wavefront, the leading edge of a convex wavefront will have
15
relatively few cells driving depolarization of many cells, shifting the balance to a greater
sink than source current. This results in reduced rate of voltage rise ahead of the
wavefront, increasing the time to sodium channel activation, effectively slowing CV. The
reverse is true in concave wavefronts, where CV is higher than that of a planar wavefront.
The curvature of the spiral wave is proposed to increase progressively towards the center,
where at the tip, the curvature achieves a critical value whereby the core becomes
inaccessible to excitatory activity. This steep wavefront slows the conduction velocity to
a critical level thus forming the PS. This allows for the meandering nature of rotors, as
the core is unexcited, but is not required to be refractory and therefore can subsequently
be excited 103. Finally, the rotor is not dependent on a fixed wavelength, as is required by
anatomic reentry and the leading circle model. In fact, electrotonus is predicted to
shorten the action potentials of cells near the core 104. This final point is of particular
importance given the previous discussion of the effect of antiarrhythmics on wavelength.
In the rotor model of fibrillation, wavelength is variable and dynamic, and thus
medications are predicted to have variable effects along the curvature of the wavefront as
the complement of currents at play will vary depending on the location.
The initiation of a rotor relies on wavebreak after the interaction of an electrical impulse
or wavefront with an obstacle 105. This process can occur in a homogeneous medium
when transient heterogeneity is introduced into the system. Classically this is performed
with an S1-S2 protocol where a first wave (S1) is followed by a second wave (S2)
oriented perpendicularly to S1 106. If S2 hits S1 before it has completely repolarized, S1
acts as a barrier to S2 propagation resulting in a rotating spiral wave. Alternatively, at
high stimulation rates, electrical and or calcium alternans can be induced where action
16
potential duration and or the amplitude of calcium transients demonstrate beat-to-beat
variation despite a constant stimulation frequency 107.
Alternans is a risk factor for cardiac arrhythmias including atrial fibrillation, and in some
cases can act as a prognostic tool for arrhythmia risk stratification and therapy 108 109 110.
In cardiomyocytes, beat-to-beat regulation of the membrane potential (Vm) and cytosolic
calcium are bi-directionally coupled and are dependent on the kinetics of APD restitution.
APD restitution refers to the dependence of an action potential on the preceding diastolic
interval to influence its duration. At rapid rates, APDs continue to shorten along their
restitution curves until they reach a critical point where the curve becomes so steep that
self-sustaining oscillations of APDs can occur 111 112 113 114. This underlying mechanism
has been hypothesized to be related to the time-dependent recovery of ion channels from
inactivation 115. Similarly, the calcium transient and conduction velocity have restitution
kinetics which have been identified as potential contributors to alternans and fibrillation
116.
1.7.2 Clinical relevance of rotors in AF
Although not universally accepted, rotor theory has become an important mechanistic
explanation for AF as well as a potential therapeutic target. Multiple clinical observations
have supported the existence of focal AF drivers in a subset of patients. These include the
ability of a single ablation lesion to terminate persistent AF that is resistant to
cardioversion 117, spatiotemporal stability and localized regions of stable frequencies
using dominant frequency mapping 118 119 120 and reproducible vectors of AF propagation
over time 121.The main limitation to developing a mechanistic understanding of rotors in
human AF however has been the resolution of current mapping systems, namely their
17
ability to differentiate the principal components of the signal from the disorganized
activation patterns. This has required significant mathematical post-processing to
identify areas of spatiotemporal reproducibility.
Recently data from a novel mapping system employing monophasic action potential
recordings and a physiologic noise filter has led to a rotor centred ablation strategy. The
Focal Impulse and Rotor Modulation (FIRM) mapping system purports to reveal stable
electrical rotors in a majority of individuals with AF. Using direct contact gold
electrodes attached to 64 pole basket catheters, both the right and the left atria are
mapped simultaneously. Electrodes are 4-6 mm apart along each spline and separated by
4-10 mm between splines 122. The software developed to interpret the tracings advertises
its ability to examine patient specific AF electrograms based upon rate-dependent
refractoriness, or as previously described, restitution, and conduction slowing. Rotors are
identified as PS and are targeted for ablation if they are stable for minutes 123. The
rationale for localized ablation lesions has been based on the theory of ablation of
microreentrant focal atrial tachycardias, for which ablation has proven successful.
The initial clinical trial, the CONFIRM trial, demonstrated a dramatic benefit in terms of
freedom from AF compared to conventional ablation (82.4% vs. 44.9%, p = 0.001) after a
median 273 days. This benefit has been shown to persist over 3-4 years follow up and
has been replicated in a multicenter experience recently reported 124 125. While there is
certainly growing data to support rotors as drivers of AF, the community of clinicians and
scientist are still mixed with respect to their belief in the ability of focal ablation lesions
to terminate AF and prevent recurrence. The hesitation is centered around the concept
that rotors are not stable, and ablation would promote anatomical re-entry and not
18
necessarily prevent recurrence of functional re-entry. There is however no published
literature to date reporting the lack of efficacy using the FIRM technology and software,
and thus critical appraisal to date has been only anecdotal. What we have certainly
learned from our growing clinical experience mapping rotors is that they can be
demonstrated using a variety of technologies and are likely important targets for therapy.
A powerful tool for the mechanistic understanding of AF and development of novel
targets would incorporate what we have learned about atrial wavelength and its response
to medications over the past century, the mathematical and clinical observations pointing
to rotors as the major drivers of AF in a model that carries the cell and tissue specific
factors already highlighted to be important in AF initiation, propagation and recurrence.
Such a tool has only recently become available with the advent and application of human
pluripotent stem cell technology. While disease modeling using human pluripotent stem
cells would not be a novel proposition, as its application and validation have
disseminated rapidly to multiple inherited and acquired diseases in multiple organ
systems including the heart, there are several unique challenges to building a model of
AF using human pluripotent stem cells. First however, one must review the source of
pluripotent stem cells, the nature of directed differentiation, and the applications of this
technology to date.
1.8 Embryonic stem cells and induced pluripotent stem cells
Embryonic stem cells (ESCs) cells are derived from totipotent cells of the developing
embryo. They are capable of unlimited proliferation while maintaining their
undifferentiated state. Human blastocyst-derived pluripotent stem cell lines were first
derived in 1998 126. Cleavage stage embryos produced by in-vitro fertilization were
19
donated by individuals after informed consent. Fourteen inner cell masses were isolated
and 5 ES cell lines derived from 5 separate embryos. These cell lines had normal
karyotypes, expressed surface markers that characterize primate embryonic stem cells,
and lacked markers defining early lineages. Specifically, the cell surface markers that
characterize undifferentiated ES cells including stage-specific embryonic antigen
(SSEA)-3, SSEA-4, TRA-1-60, TRA-1-81 were expressed, while the cell lines did not
stain for SSEA-1. After 5 months in culture, these cells retained their ability to derive all
three embryonic germ layers; endoderm, mesoderm and ectoderm. When injected into
severe combined immune-deficient (SCID)-beige mice, each injected mouse formed a
teratoma that included gut epithelium (endoderm), cartilage, bone, smooth muscle and
striated muscle (mesoderm) and neural epithelium, embryonic ganglia and stratified
squamous epithelium (ectoderm). When grown in vitro to confluence, ES cells
differentiated spontaneously.
Since the introduction of human stem cell lines, a second source for pluripotent stem cells
has arisen. In 2006, Takahashi and Yamanka demonstrated that adult skin fibroblasts in
mice can be reprogrammed into ES like cells by using a retroviral vector to force the
expression of 4 transcription factors: Oct3/4, Sox2, Klf-4 and c-Myc 127. Subsequent
studies were able to recapitulate these findings using the same 4 transcription factors in
human fibroblasts and showed that these cells, human induced pluripotent stem cells
(hiPSCs), were capable of self-renewal, could remain undifferentiated in culture, and
could give rise to all somatic cell types 128. The potential to generate an unlimited number
of any of the hundreds of cell types in the human body has innumerable applications in
the fields of biomedical research and regenerative medicine. The potential however
20
cannot be actualized without the ability to control and drive the directed differentiation
into the cell type required.
1.8.1 Directed differentiation
Three basic differentiation methods have been developed and applied for the
differentiation of ESCs; the formation of 3-D aggregates known as embryoid bodies
(EBs), the culture of ESCs as cell sheets, and the culture of ESCs directly on stromal
layers 129. While early protocols using fetal calf serum were difficult to reproduce and
were generally poorly optimized for the generation of a single cell lineage, several
advances including the use of serum-free media with specific inducers, and the
development of reporter ESCs to monitor progression through the developmental and
differentiation steps have advanced the field 130 131 132 133 134 135 136 137. With these new
tools, it became possible to approach the problem of directed differentiation through the
lens of developmental biology, applying insights from other model systems towards the
goal of recapitulating human developmental milestones in vitro. For the purpose of this
discussion, the focus will be primarily on the EB culture method of generating
cardiomyocytes for disease modeling and regeneration.
Understanding the steps taken by early cardiac progenitors in the developing mouse
embryo has assisted in the identification of critical signalling pathways, gene expression
profiles and surface markers that direct in vitro differentiation of hESCs. Perhaps the
most important event in embryogenesis lies in gastrulation, when uncommitted epiblast
cells migrate through a transient structure known as the primitive streak and exit as either
mesoderm or definitive endoderm. Brachyury, a T-box transcription factor, is up-
regulated as the epiblast cells enter the primitive streak, and then rapidly down regulated
21
as the newly formed mesodermal cells exit the primitive streak. Specification of distinct
subpopulations is controlled both temporally and spatially suggesting different signalling
environments responsible for lineage specification. Members of the TGFβ family
including BMP4 and Nodal as well as the Wnt family are essential pathways during these
developmental steps 138 139 140. These agonists work in concert with regional expression
of inhibitors to create domains that are conducive to specifying germ layer induction 141.
Subsequent studies demonstrated that the same signalling pathways can be manipulated
in vitro to regulate primitive streak development, setting the stage for germ layer
induction 129.
The early stages of mesoderm induction can be monitored by the up regulation of fetal
liver kinase-1 (Flk-1) and PDGF receptors 142,143. The induction of early cardiac
progenitors requires initially Wnt signalling and then its subsequent inhibition in order to
specify the mesoderm, while using various concentrations of BMP4 and activin may
generate different subpopulations 144,145. Cardiomyocytes are derived from the lateral
plate mesoderm and in two waves of development of cells marked by the VEGF receptor
Flk-1 and the transcription factor Nkx2.5 146,147. Mesodermal cells migrate to the anterior
region of the embryo and organize into the cardiac crescent 148. Ultimately the cardiac
crescent fuses to become the primitive heart tube which subsequently undergoes a
complicated series of changes, or looping, resulting in heart chamber formation. The
differentiation protocols have improved tremendously over the last decade resulting in
cardiomyocyte yields of 80-90% 149.
The study of the electrophysiologic properties of ES derived cardiomyocytes has given us
significant insights into their developmental stage, and has become an important
22
functional assay in the search of differentiating cells towards further cell type specificity.
While many aspects of the human pluripotent stem cell (hPSC) derived cardiomyocytes
are comparable to adult cardiomyocytes, there are several notable differences. Principally
it is believed that these differences are attributable to the immature nature of these cells,
as they beat spontaneously, have unorganized sarcomeres, and have a different
complement of ion channels compared to adults 150. In particular, the excitation-
contraction machinery appears to be immature with hPSCs lacking clear T-tubules, and
displaying disorganized sarcomeric striations and immature Ca2+ handling 151,152. Time
has been shown to be an important aspect of maturation, with cells cultured for extended
periods of time acquiring modifications in current densities and properties, increased
expression of a “more mature” host of ion channels, and significant improvements in cell
morphology 153. No studies to date however have been able to demonstrate a dramatic
impact on the relatively immature AP shape, or the Ik1 expression felt to be at least in part
responsible for the relatively depolarized resting membrane potential compared to adult
cardiomyocytes 150.
We have also learned that standard protocols154,155to generate CMs from hESCs produced
mixed populations of atrial-, pacemaker- and ventricular-like cardiomyocytes. The
majority of cells derived carry a ventricular like action potential (AP) morphology (60-
90%), with a minority of cells exhibiting atrial or nodal like AP morphologies 156. There
has been a substantial amount of evidence, both in vivo and in vitro, implicating retinoic
acid (RA) signalling as an important player in atrial specification.
23
1.8.2 Retinoic acid signaling
RA signalling has been shown to regulate anterior-posterior polarisation of the heart 157.
RA treatment of mouse and chicken embryos leads to oversized atria and smaller or
missing ventricles, while the opposite is true for embryos in which RA signalling is
inhibited at critical time points 158 159. Furthermore, studies of mouse embryonic stem
cells indicated that RA signalling promotes the expression of atrial-specific genes.
Finally, exogenous treatment of RA has been shown to direct the differentiation of hESCs
into atrial like myocytes 160.
RA is synthesized from vitamin A (retinol) through a chain of 2 oxidative sequences. In
the second step, which is thought to be rate limiting, all-trans retinaldehyde is converted
to all-trans RA by any of three related aldehyde dehydrogenases, Raldh1, Raldh2, and
Raldh3. Previous studies have implicated Raldh2 as the main Raldh involved in early
cardiac development 161. Our group hypothesizes that early cardiac progenitors that
express Raldh2 are programmed to synthesize and thus respond to RA, and that atrial
precursors and not ventricular precursors would thus express Raldh2. Since RA readily
diffuses across cell membranes, ventricular progenitors would have to “protect”
themselves from RA being synthesized by neighbouring cells, and could do so by
expressing an enzyme that degrades RA. Cyp26a1, a member of the Cyp26 family of
cytochrome p450 enzymes, converts RA to metabolites that are not bioactive.
Upregulation of Cyp26a1 has been shown to attenuate RA signaling in the prospective
rostral spinal cord of zebra fish, thus limiting the downstream signaling involved in the
expression of hox genes, and demonstrating its functional role in determining the
hindbrain-spinal cord boundary 162.
24
The potential therefore exists to drive cardiac differentiation to specific cell fates as
would be required for disease modeling of atrial tissue. The validation of the successful
generation of a specific cell type will require multiple modalities in order to verify the
cell’s molecular and physiologic fingerprint. The tools that are currently relied on most
heavily are the quantitative reverse transcriptase polymerase chain reaction (qRT-PCR)
targeted analysis of RNA expression, flow cytometry, immunostaining and patch
clamping. Important for the purposes of this study will be the ability to differentiate the
two major cell lineages of cardiomyocytes in our cultures, atrial like cells and ventricular
like cells.
1.8.3 Molecular markers of atrial and ventricular cardiomyocytes
Recently, comparative proteomic and transcriptomic analysis of human fetal atria and
ventricular samples has been performed identifying and validating chamber enriched and
chamber specific proteins 163. Our group has focused on the following atrial enriched
markers: atrial natriuretic factor (ANF), a potassium inward rectifying channel (KCNJ3),
Connexin-40 (Cx-40), and the alpha 1D subunit of the L-type voltage-dependent calcium
channel (CaV1.3). ANF is encoded for by the gene natriuretic peptide A (NPPA) and
belongs to a family of natriuretic peptides involved in fluid and electrolyte homeostasis.
Mutations in NPPA have been associated with familial AF. KCNJ3 or Kir3.1, associates
with other G-protein-activated potassium channels to form a heterotetrameric pore
forming complex. These multimeric G-protein gated inwardly rectifying potassium
(GIRK) channels are important in atrial electrophysiology and can be activated by
muscarinic M2 receptors. Cx-40 is encoded for by the gene GJA5. Connexins form gap
junctions in the heart through which charged ions flow between neighbouring cells thus
25
facilitating action potential propagation. The two most highly expressed connexins in the
heart are Cx-40 and Cx43, however Cx-40 is expressed in atrial cardiomyocytes and not
ventricular cardiomyocytes, and has been associated with a heritable form of AF. CaV1.3
is highly expressed in atrial compared to ventricular tissue, contributes significantly to the
repolarization process in human atria, and has been linked through genome wide
association studies to AF in humans 164.
In a similar fashion, our group has utilized the absence of ventricular markers myosin
regulatory light chain-2 (MYL2) and Iroquois-class homeodomain protein 4 (IRX4) to
mark atrial cardiomyocytes and their precursors. Ventricular myosin light chain-2 (MLC-
2v) refers to the ventricular form which plays a critical role in embryonic cardiac
development and function as well as representing one of the earliest markers of
ventricular specification 165. IRX4 expression is restricted to the ventricular precursors
during the embryonic formation of the linear heart tube, and is absent from both atrial and
outflow tract precursors 166.
1.9 Disease modeling using hPSCs
The careful dissection and directed differentiation of cardiomyocyte subtypes is a
relatively new and important step forward in the field of disease modeling using hPSCs.
Prior to this advance, investigators have generated predominantly ventricular like cells
and thus focused on diseases of the ventricles. The first patient specific disease model to
be studied in depth was the long QT syndrome in 2010 when hiPSC derived
cardiomyocytes were clearly shown to have the capacity to recapitulate the clinical
disease in a dish. Since then there has been an explosion of interest and innovation
around this novel technology167 and its application in the study of innumerable cardiac
26
diseases from inherited arrhythmia syndromes to cardiomyopathies168-170. Disease
modeling with hPSCs has become an important part of the armamentarium for gene
discovery, its subsequent functional analysis, and drug testing. For the purpose of this
discussion, the lessons learned in modeling congenital and acquired long QT syndrome
will be reviewed in further detail, as they will act as the building blocks upon which this
thesis was designed and carried out.
1.9.1 cLQTS
The congenital long QT syndrome (cLQTS) is a life threatening disease that represents a
leading cause of sudden cardiac death in the young 171. The electrocardiographic features
of the disease are QTc prolongation and T wave abnormalities at rest, and failure of the
QTc to shorten with exercise and epinephrine 172. The QT interval represents the
depolarization and repolarization phases of the cardiac action potential, as previously
reviewed. Decreases in repolarizing outward potassium currents or increases in
depolarizing inward sodium or calcium currents can lead to prolongation of the QT
interval, and not surprisingly then, mutations in genes encoding ion channels have been
identified as the most common pathogenic variants. Approximately 1 in 2500 healthy
live births will have an abnormally long QT interval and have congenital LQTS,
transmitted via an autosomal dominant inheritance pattern 171.
Beta-blocker therapy is the primary treatment for most patients with LQTS and offers
substantial protection from fatal cardiac events 173-176. Patients who are intolerant or
refractory to beta-blockers can be offered left cardiac sympathetic denervation 177.
Patients who have cardiac events while on beta-blockers, who have suffered a cardiac
arrest, or who are deemed sufficiently high risk, can be offered an implantable
27
cardioverter defibrillator (ICD)178-181. ICD therapy however has lifelong implications, and
complications are common especially in young patients who may have the device for >20
years. In a recent review of an academic tertiary center’s outcomes for primary
prevention ICDs, 35% of patients had ICD related morbidity, and none of their LQTS
patients received appropriate shocks for LQTS related arrhythmias 182.
The majority of efforts at improving risk stratification in patients with LQTS have
focused on our rapidly improving understanding of genetics. Consequently LQTS has
become one of the best-understood and characterized monogenic genetic diseases,
serving as a model for the investigation of genotype-phenotype interactions. This
mechanistic basis and understanding of disease has afforded clinicians an improved
patient specific management strategy. This has also given researchers a unique depth to
the understanding of a monogenic disease that often carries a clear, distinct, and easily
measurable phenotype. Mutations in ion channel genes causing QT interval prolongation
do so by prolonging the action potential duration. Delayed repolarization can facilitate
early after depolarizations (EADs), a process that has been proposed to relate to L-type
calcium channel reactivation, and is dependent not only on time but also on membrane
potential 183 184. It is however widely accepted that torsades-de-pointes (TdP) is triggered
by EADs under conditions of a prolonged QT interval 185. Drugs that further prolong the
APD can aggravate the underlying QT prolongation and put patients at risk for sudden
cardiac death.
The first inherited cardiac disorder to be modeled was a patient with type-2 LQTS due to
a missense mutation in the KCNH2 gene affecting the pore-forming region of the HERG
channel 186. This mutation leads to a significant reduction of the rapid component of the
28
delayed rectifier potassium current (Ikr). After an iPS line was generated and
differentiated into beating cardiomyocytes, intracellular recordings revealed APD
prolongation compared to control cells as well as a significant reduction in Ikr current. A
significant proportion (66%) of LQTS iPSC derived cardiomyocytes displayed EADs,
whereas control cells did not. The additional stress of specific Ikr blockers prolonged the
APDs further and incited an increased number and complexity of EADs. Application of
potential therapeutic agents including calcium channel blockers and pinacidil, a KATP-
channel opener, resulted in APD abbreviation and elimination of EADs.
The observations in this iPS model of LQT2 are striking, however the implications are
limited by the investigation of a tissue based arrhythmia at a single cell level. Under
normal physiologic conditions, there is dispersion across the ventricles during
repolarization 187. While the cellular basis for this continues to be debated, one theory that
has gained popularity relies on the gradient of cell types across the myocardial wall.
From epicardium to myocardium, and possibly M cells in between, ventricular
cardiomyocytes differ in terms of their repolarization properties. The ionic determinants
of the differing properties of these cell types has also been proposed to underlie the
enhanced transmural dispersion of repolarization (TDR) under conditions of prolonged
QT interval that can predispose patients to TdP and sudden cardiac death. An increase in
TDR may be essential for the development of TdP, serving as a functional reentrant
substrate for its maintenance, as well as critically facilitating the propagation of an EAD
to generate the first initiating beat 188 189 . In the absence of TDR in canine and rabbit
ventricular wedge preparation models, TdP does not develop in the setting of QT
prolongation despite frequent phase 2 EADs. Further evidence dissecting the surrogate
29
outcome of QT prolongation and EADs from the clinically relevant outcome of TdP is
the observation that the incidence of TdP is not proportional to the extent of QT
prolongation in drug-induced or acquired LQTS 185. Cardiologists have long since
learned the perils of treatment paradigms based on surrogate outcomes. Perhaps the most
notable lesson has come from the Cardiac Arrhythmia Suppression Trial (CAST),
designed on the premise that suppression of ventricular ectopy after a myocardial
infarction reduces the incidence of sudden cardiac death 190. Based on a well documented
observation that ventricular ectopy was associated with sudden death after myocardial
infarction, antiarrhythmics were commonly prescribed to suppress them until the CAST
trial definitively showed that this management strategy was associated with increased
mortality. Despite the clear limitations of single cell physiology studies and surrogate
outcomes, we have gained significant insights into the utility of hPSC technology for
disease modeling and drug screening from recent studies of congenital and acquired
LQTS.
1.9.2 acquired LQTS
New drug development suffers from low phase II clinical success rates, and a daunting
price tag of 1.8 billion dollars and 12 years of development, testing, and regulatory
processes spanning the time from discovery to commercial launch 191. The need for new
technology is highlighted further by the fact that the attrition rate of new therapeutics for
any disease condition is 89% after preclinical animal testing, with 1/3 being ineffective
and an additional 1/3 having safety issues192, with cardiac toxicity being the major factor
in this attrition193. Much of this attrition rate has been attributed to the poor predictive
power of animal models. Several high profile drug withdrawals have led to mandatory
30
pre-clinical screening policies to detect compounds which may prolong the QT interval
and predispose patients to TdP 194. So called acquired QTc prolongation is much more
common than congenital and typically ascribed to drugs that incidentally block the IKr
channel, but can exert this effect through any number of mechanisms, and thus prolong
the ventricular action potential duration 195,196. Current FDA mandated methods rely on
cell lines artificially expressing single cardiac ion channels, or whole animal heart and
purkinje assays 197. These screens have obvious and important limitations, with poor
sensitivity and specificity profiles for human translation, and a host of untestable
electrophysiologic toxicities that are not considered 194. Pluripotent stem cell assays have
been shown to act as efficient, reproducible screens for the efficacy of new compounds as
well as the safety profile of compounds in development 198. With the advent and
explosion of genetic engineering techniques, hPSCs carry the unique privilege of being
the only human model system that can incorporate the breadth of genetic diversity of
patient populations, while also holding the promise to personalize medical therapy and
risk stratification at the level of a single patient.
The first major effort towards the development of hPSC based drug screening was
performed at the single cell level in hiPSCs derived from healthy subjects and patients
considered to be at higher risk of developing drug induced QT prolongation. The high
risk population drew from cohorts of patients with congenital long QT syndrome, familial
hypertrophic cardiomyopathy and familial dilated cardiomyopathy 197. Cardiomyocytes
generated from high risk hiPSCs demonstrated increased susceptibility to drug induced
APD prolongation, early and delayed after depolarizations compared to control.
Subsequently, investigators from the same group generated a library of genome edited
31
iPSC derived cardiomyocytes, with mutations again representing the highest risk patient
populations, and compared the effects of drugs to isogenic controls. The application of
genetically modified human cell lines demonstrated clear utility in efficient high
throughput screens for small molecules and chemical compound testing 198. Of course,
the same limitation ascribed to the LQT2 model previously described applied to single
cell drug screening assays, screening for a surrogate markers (EADs) of the actual feared
clinical outcome (TdP).
The atria and the ventricles share the same fundamental electrical properties at a single
cell level that impact the action potential profile, and at a tissue level when considering
electrical wave propagation and repolarization. Accordingly similar principles apply to
understanding arrhythmias in the two chamber types, specifically, the determinants of
wavelength, and the clinical patterns of arrhythmogenesis including fibrillation involving
reentry, conduction block, and spiral wave formation (i.e. rotors). However, there are key
differences in the physiology and pathophysiology of atria and ventricles, principally
their electrophysiologic signatures and their cell to cell communication, that can be
exploited for therapy, and tested for safety. Unfortunately, our understanding of the
mechanisms of atrial arrhythmias remains dismal and treatment approaches are grossly
ineffective and unsafe 199. One of the major barriers in the pipeline of drug development
has been the lack of appropriate models. The challenges of interspecies differences in
receptor subtypes, distribution and signaling200 are compounded by significant
differences in the ion channels and gap junctions201,202 that work in concert to dictate the
physiology and pathophysiology involved in human AF 102. The major objective of our
studies is to develop a novel platform, using human pluripotent stem cells (hPSCs), to
32
study the pathophysiology of AF in order to identify novel targets for therapy and as a
platform to study the mechanisms required for proper treatment and prevention of AF.
33
Chapter 2
Hypotheses and Research Aims
This study proposed to generate an in vitro model of AF using human pluripotent stem
cells. Building on the work to date from the Keller lab and published literature, we
generated cardiomyocytes from hESCs that carry the molecular and electrophysiologic
hallmarks of atrial cells. The differentiation scheme developed relies principally on
retinoic acid signaling in driving the atrial cell fate, as well as fine tuning of the BMP and
Activin signaling pathways in order to generate cells that mimic early atrial precursors.
Once validated at the single cell level, using hESC derived ventricular-like cells as
controls, we have moved forward in generating a simple tissue model in the form of a
multicellular sheet. The atrial chambers are thin compared to the ventricles, with a mean
thickness of 1-2 mm compared to ~15 mm, and thus multicellular sheets may be a
reasonable approximation of atrial tissue. This platform has the potential to facilitate
optical mapping of the electrophysiologic properties of the atria under physiologic and
pathophysiologic conditions. Thus our hypothesis is that multicellular sheets of hESC
derived atrial cardiomyocytes will serve as a reliable model of human AF and has the
potential to act as a tool to understand the mechanism of action of anti-arrhythmic drugs,
and to study and screen for novel therapeutics. Based on this hypothesis, the objectives
of this study are the following:
1. Generate atrial like cardiomyocytes from hESCs through the manipulation of Activin
and BMP signaling pathways and the addition of retinoic acid
2. Generate atrial like tissue in the form of cell sheets
3. Induce atrial fibrillation in the cell sheets
34
4. Develop on optical mapping setup, and the requisite signal processing software,
capable of monitoring the electrophysiologic properties of cell sheets during sinus rhythm
and AF
6. Test the effects of AADs that are commonly used for the treatment of human AF on
the cell sheets using the acquisition setup and signal processing
35
Chapter 3
Materials and Methods
3.1 Generation of cardiomyocytes from hESCs
HES3 NKX2-5egfp/w cells were maintained on irradiated mouse embryonic feeder cells in
hESC media consisting of DMEM/F12 (50:50;MEdiatech, Hedndon, VA) supplemented
with 20% knock-out serum replacement (SR), 100 μM nonessential amino acids, 2mM
glutamine, 50U/mL penicillin, 50 μg/ml streptomycin (Invitrogen Grand Island, NY), 10-
4 M Β-mercaptoethanol (Sigma, St Louis, MO), and 20 ng/mL hbFGF (R&D Systems,
Minneapolis, MN) in 6-well tissue culture plates 203. Cells were passaged to new feeders
as single-cell suspensions following dissociation with TrypLE (Life technologies).
Embryoid bodies for differentiation were generated after feeder depletion and growth of
ES cell colonies to 80% confluence. Single cell suspensions were formed after
dissociation with TrypLE. Embryoid bodies were formed by plating small aggregates at a
concentration of 500,000 cells/ml in 2 ml basic media containing StemPro34 (Invitrogen),
2mM glutamine, 4 X 10-4M monothioglycerol (MTG), 50 μg/ml ascorbic acid (Sigma),
and 0.1 ng/ml BMP-4 (R%D Systems). The aggregates were incubated at 37°C in a
hypoxic environment of 5% CO2, 5% O2, and 90% N2 on a rotator set at 70 rpm for 24
hours. Following this, the aggregates were spun in a centrifuge at 300 rpm, washed using
IMDM, and then re-suspended in the induction media.
The following concentrations of factors were used for the ventricular (control)
differentiation protocol: BMP-4, 10 ng/ml; human bFGF, 5 ng/ml; activin A, 6ng/ml;
IWP2,0.5 ng/ml; and human VEGF, 10 ng/ml. This is referred to subsequently as
10B/6A, control, or “ventricular” differentiation protocol. The factors were added with
36
the following sequence: days 1–3, BMP4, bFGF and Activin A; days 3–5, VEGF and
IWP2; day5-12 VEGF, and days 12-20 only backbone media as previously described.
Media changes were performed by centrifuging the EBs at 800 rpm and washing the EBs
with IMDM. Cultures were maintained in a hypoxic 5% CO2/5% O2/90% N2
environment for the first 12 days and then transferred to a normoxic 5% CO2/air
environment.
The atrial differentiation protocol was carried out as above, with the following changes;
BMP-4 and Activin were titrated down to 3 ng/ml and 2 ng/ml respectively. On day 3,
retinoic acid (RA) was added to the culture at a concentration of 0.5 ng/ml. Other than the
addition of RA, this protocol only differed from the control protocol in terms of the
concentrations of BMP4 and Activin used at T3 employing lower concentrations. This is
referred to in text and figures as 3B/2A +RA, or “atrial” differentiation protocol.
3.2 Flow Cytometry and Cell Sorting
For cell-surface antigens, staining was carried out in PBS with 3% FCS. For intracellular
antigens, staining was carried out on cells fixed with 4% paraformaldehyde in PBS.
Staining was done in PBS with 3% FCS and 0.5% saponin (Sigma). Cells were stained at
a concentration of 2.5 × 106 cells/ml with anti-KDR- APC (R&D Systems; 1:10) and
anti-PDGFRA– PE (R&D Systems; 1:20), anti-SIRPA–PE-Cy7 (clone SE5A5;
BioLegend; 1:500), anti-CD90-APC (BD Pharmingen, 1:2000anti-CTNT (clone 13-11;
Thermo NeoMarkers; 1:400), goat anti-mouse IgG–APC(BD; 1:200). Cells assayed for
Aldefluor (STEMCELL Technologies) were prepared based on manufacturers
instructions. Incubation with the Aldefluor reagent was 45 minutes. Stained cells were
analyzed on an LSRII flow cytometer (BD Biosciences). For FACS, the cells were sorted
37
at a concentration of 106 cells/ml in IMDM/5% FCS using a FACSAriaTMII (BD
Biosciences) cell sorter (SickKids-UHN Flow Cytometry Facility). Data were analyzed
using FlowJo software (Treestar).
3.3 Molecular Markers
3.3.1 Quantitative Real-Time PCR (qRT-PCR)
Total RNA was prepared with the RNAqueous-Micro Kit (Ambion) and treated with
RNase-free DNase (Ambion). RNA (500 ng to 1 μg) was reverse transcribed into cDNA
using random hexamers and Oligo(dT) with Superscript III Reverse Transcriptase
(Invitrogen). qRT-PCR was performed on a MasterCycler EP RealPlex (Eppendorf) using
QuantiFast SYBR Green PCR Kit (Qiagen). Expression levels were normalized to the
housekeeping gene TATA binding protein (TBP).
3.3.2 Immunostaining
Immunostaining was performed using the following antibodies: Polyclonal rabbit anti-
MLC2V (Synaptic systems, 1:1000) with the secondary antibody anti-rabbit rabbit IgG
Cy3 (Jackson ImmunoResearch), and mouse anti-CTNT (Thermo NeoMarkers; 1:100)
with the secondary antibody anti-mouse IgG Alexa647 (Invitrogen). DAPI (Life
Technologies, SlowFade Gold) was used to counterstain nuclei. The stained cells were
visualized using a fluorescence microscope (Leica CTR6000) and images captured using
the Leica Application Suite software.
3.4 Cardiac Electrophysiology
3.4.1 Single Cell Electrophysiology
Spin EBs resulting from control and RA-treated differentiations were dissociated at day
20 to single cells using type B collagenase and TrypLE. They were sorted based on
38
DAPI-/SIRPA+/GFP+/CD90- surface markers as previously described, generating a pure
population of cardiomyocytes. Single cells were plated on matrigel-coated coverslips.
Electrophysiological measurements were performed 7-10 days after dissociation and
plating.
Spontaneous action potentials (APs) were recorded using the patch clamp technique in
the whole cell configuration 204..
APs were recorded at room temperature (22-23 °C), and
were not corrected for the calculated liquid junction potential 205.
Pipettes were pulled
from borosilicate glass (with filament 1.5 mm OD, 0.75 mm ID, Sutter Instrument
Company) using a Flaming/Brown pipette puller (model p-87, Sutter Instrument
Company) and heat polished. The resistance of these pipettes was 8-10 MΩ when filled
with recording solution. Micropipettes were positioned with a micromanipulator
(Burleigh PCS-5000 system) mounted on the stage of an inverted microscope (Olympus
IX70). Seal resistance was 2-15 GΩ. Myocytes were placed into and perfused with a bath
solution of Tyrode’s containing (in mmol/L) 140 NaCl, 2.5 KCl, 1.5 CaCl2, 1 MgCl2. 10
HEPES and 5.5 D-glucose with pH adjusted with NaOH to 7.4. Standard AP internal
solution was used to fill the pipettes which consisted of (in mM) 100 K aspartate, 10 KCl,
3 NaCl, 5 NaHCO3, 3 KHPO4, 1 MgCl2, 5 MgATP, 10 HEPES, 75 μM EGTA, with pH
adjusted to 7.2 with KOH.
Membrane potential was controlled with an Axon headstage (CV 203BU) connected to
an Axopatch 200B voltage-clamp amplifier (Axon Instruments, Foster City, CA). Data
was digitized (Axon Digidata 1320A) and acquired using Axon Clampex software
(pClamp version 8.2.0.232). Prior to cell attachment and formation of a GΩ seal,
electrode potential was adjusted to a baseline of zero current. An agar-salt bridge was
39
used as the reference electrode. Rupture of the cell membrane patch was achieved using
pressure disruption by brief suction. Following rupture of the cell membrane, the
membrane capacitive transient was elicited by small depolarizing voltage pulses. Data
was analyzed using Clampfit (Molecular Devices, Sunnyvale, CA, U.S.A).
3.4.2 Micro Electrode Array (MEA) electrophysiology
MEA chips were cleaned, autoclaved, and coated with dilute matrigel overnight at 37 °C.
hESC-CMs were digested using collagenase B at a concentration of 250 U/ml in Hank’s
balanced solution (Life technologies) overnight at a concentration of 5 million cells/ml.
TrypLE was used to create single cell suspensions which were sorted based on DAPI-
/SIRPA+/NKX2-5 GFP+/CD90- surface markers as previously described, generating a
pure population of cardiomyocytes. Single cells were plated on standard 60 electrode
MEAs at a concentration of 1-2 million cells per 150 μL drop in backbone media plus 1
ng/ml ROCK inhibitor (Y-27632 Dihydrochloride Hydrate,Toronto Research Chemicals,
cat.no Y100500) generating cell sheets that were 1-1.5 cm in diameter. Extracellular
recording was performed using a MEA1060INV MEA amplifier (Multi Channel Systems,
Reutlingen, Germany). Output signals were digitized at 10 kHz by use of a PC equipped
with a MC-card data acquisition board (Multi Channel Systems). Standard measurements
were performed in IMDM supplemented with NaCl (final concentrations in mmol/L: 140
NaCl, 3.6 KCL, 1.2 CaCl2, 1 MgCl2. 10 HEPES and 5.5 D-glucose). During recordings,
temperature was kept at 37 °C. Data were recorded using Cardio2D+ (Multi Channel
Systems,) and analyzed off-line with Cardio2D (Multi Channel Systems).
40
3.4.3 Optical Mapping Electrophysiology
150 μL of 100% concentrated matrigel drops were placed in the centre of 35 mm tissue
culture treated petri dishes (Falcon). The dishes were kept on ice for 30 minutes and then
the matrigel was removed leaving a thin coat of matrigel with a diameter of 1 cm. The
petri dishes were incubated at 37 °C overnight. hESC-CMs were digested using
collagenase B at a concentration of 250 U/ml in Hank’s balanced solution (Life
technologies) overnight at a concentration of 5 million cells/ml. TrypLE was used to
make single cell suspensions which were sorted based on DAPI-/SIRPA+/NKX2-5
GFP+/CD90- surface markers as previously described, generating a pure population of
cardiomyocytes. Single cells were plated on the matrigel coated petri dishes at a
concentration of 1.25 million cells per 150 μL drop in backbone media plus 1 ng/ml
ROCK inhibitor (Y-27632 Dihydrochloride Hydrate,Toronto Research Chemicals, cat.no
Y100500) generating cell sheets of 1 cm diameter. To improve the stability of the cell
sheetsand maintain confluence, various percentages of CD90+ cells generated from the
sort were spiked into the cell sheet, from 5-10%. Cell sheets were incubated under
normoxic conditions at 37 °C for 2-3 weeks prior to imaging. Once the differentiation
was optimized to generate 80-90% cardiomyocytes consistently, cells were digested into
single cells, using the aforementioned techniques, and plated directly on the matrigel as
cell sheets without sorting. This ultimately had the effect of generating confluent cell
sheets with greater consistency.
Optical mapping was performed using the voltage sensitive dye
AminoNaphthylEthenylPyridinium (Di-4-ANEPPS)(Life Technologies). Di-4-ANEPPS
is an ampiphilic compound with two hydrocarbon chains that allow it to anchor into the
41
plasma membrane. When bound to the membrane, the chromophore aligns
perpendicularly to the membrane/aqueous interface and undergoes reduction in emission
intensity(~10% per 100 mV depolarization) when excited at 540 nm and monitored at
680 nm emission206. The incubation was optimal when cold Di-4-ANEPSS (on ice) was
applied to the cell sheets and then incubated for 30 minutes at 10 μM. Blebbistatin, an
inhibitor of the adenosine triphosphatases (ATPases) associated with class II myosin
isoforms in an actin-detached state, was employed to stop cardiomyocyte contraction in
an effort to avoid motion artifact. Blebbistatin has high specificity for Myosin II and has
been demonstrated previously to be an effective excitation-contraction uncoupler,
immobilizing cardiac preparations without affecting AP morphology and intracellular
calcium handling207. Blebbistatin (Sigma) was used at a concentration of 10 μM.
Imaging was performed in IMDM supplemented with NaCl (final concentrations in
mmol/L: 140 NaCl, 3.6 KCL, 1.2 CaCl2, 1 MgCl2. 10 HEPES and 5.5 D-glucose). The
Di-4-ANEPPS and blebbistatin were washed and replaced with the same IMDM
containing 2 μM Di-4-ANEPPS. The petri dish was placed on a plate warmer
constructed using copper piping sandwiched by 2 aluminum plates. The copper piping
was connected to a water bath heater and pump with rubber tubing. Sylgard was used to
coat the aluminum plate where a well had been created and sized for the 35mm petri dish.
This generated even heating across the medium in the petri dish to a temperature of 36-
37°C.
The tissue was illuminated using a mercury light source (X-Cite Exacte, Lumen
Dynamics, Mississauga, ON, Canada) with a 525 ± 50 nm band-pass filter. Fluorescent
light was collected using a 645 ± 75 nm band-pass filter. Images were captured using an
42
electron multiplying charge coupled device (EMCCD) camera (Cascade 128+, Cascade
Evolve, Photometric, Tucson, AZ, U.S.A) connected to an Olympus MVX-10 upright
microscope (Center Valley, PA, U.S.A) equipped with a 0.63x c-mount adapter and a
0.38x lens relay. Frames were captured at 1408 fps, at 2x2 binning and 64x32 pixels
resolution, or 522 fps at 1X1 binning, using Image Pro Plus (Media Cybernetics,
Rockville, MD, U.S.A) software. Stimulation and sensing electrodes were constructed
from platinum wire coated with Sylgard for electrical insulation. Stimulation electrodes
were designed in a bipolar configuration in order to achieve point stimulation. These
electrodes were connected to Pulsar 6i & 6b stimulators (FHC Inc, Bowdoinham, ME,
U.S.A) for programmed stimulation. Capture threshold was determined either by direct
visualization of cell sheet contraction, or simultaneous optical mapping of signals to
confirm rate. We were able to obtain 100% capture on the majority of cell sheets using
40mV with a pulse duration of 2 ms. Burst (“on and off”) pacing at a cycle length of 50
msec for 1-2 minutes was used for AF induction. AF induction was defined by rotor
initiation in a cell sheet that had demonstrated nodal activation at baseline. Sequences of
2048 frames were captured serially to maximize data acquisition while minimizing
photobleaching.
3.4.3.1 Signal Processing and Data
Optical action potential duration (OAPD) measurements were performed using ImageJ,
an open source image processing program designed for scientific multidimensional
images. Images were analyzed using minimal processing as defined as an averaging of
the image sequences which were then appended, and the unprocessed images were
subtracted from the processed images. This processing, in addition to removing the
43
background, enhanced the pixel intensity changes associated with depolarization to aid in
analysis. In other words, for every pixel, let I(x,y) be the input. Let O1 be the first stage
of output. Then O1(x,y) = Iave(x,y) - I(x,y) + 10000, where Iave(x,y) is the time average of
pixel(x,y) and I(x,y) is the intensity of the individual pixel.
Optical APs were generated by measuring stacks over regions of interest (ROIs) using
both raw data and minimal processing as previously defined. When recordings had
suitable signal to noise ratios, 1 pixel was used to generate optical APs. When the signal
to noise ratio was less favourable, ROIs were selected, up to an area of 10 pixels, over a
region of the cell sheet that was in a single phase, meaning the activation wavefront
would cross the ROI at the same time in each pixel included.
In order to generate activation maps and conduction velocity maps, recordings were
analyzed using Scroll software (courtesy of Sergey Mironov). Processed images were
subjected to temporal and spatial filtering to reduce high frequency noise. The activation
time (the time of the first derivative of the flurorescence signal corresponding to the
steepest segment of the optical AP) was defined for each pixel first, followed by
generation of the activation map and the isochronal activation map. The activation time
delay between two adjacent neighbours in the vertical, or horizontal components divided
by the distance between the two neighbours was used to calculate the vector gradient for
each pixel. The resultant local gradient vector was then reciprocated to obtain the
conduction velocity vector magnitude, and inversed to gain conduction velocity vector
direction. The resultant vector velocity map was then used to guide ROI analysis of
conduction velocity. The cell sheet was divided into tertiles based on their distance from
the focus of electrical generation and propagation in the cell sheet (either nodal or the
44
centre of a rotor, identified as the phase singularity). ROIs were taken across each tertile,
and then combined to generate a mean and standard deviation of conduction velocity
across the cell sheet.
Generating APD maps: Advanced Signal Processing
Due to the low S/N ratios, further processing was required. The first filtering was spatial.
Either of the following were used: 7x7 kernel averaging or bandwidth filtering using a 2D
FFT. The 2D FFT implementation was that given by ImageJ. The bandwidth used was 8 -
64 pixels. Let O2 be the 2nd output stage.
O2(t) = spatial filtering(O1(t))
A temporal filtering was also required. Pixel-by-pixel, this was done by taking the FFT of
O2. Zeroing the coefficients associated with higher frequencies k as well as the negative
frequencies, and then reconstructing the signal with an IFFT. The k = Cutoff frequency
was chosen to be as high as possible that will enable a good estimation of the action
potential duration. The cutoff frequency ranged from 5Hz to 58Hz. In all cases, this
meant including the fundamental frequency and several of its harmonics to many of its
harmonics. Let O3 be the third output stage.
O3(x,y) = IFFT(0 to cutoff frequency of O2(x,y))
Depending on the amount and type of filtering the baseline of O3 may have a low
frequency component. A filtering is performed by using the rolling ball algorithm to
establish the baseline. In this variation of the rolling ball algorithm, the height of the ball
is scaled to the signal swing O3 as the ball is moved temporally beneath the signal. The
radius of the ball is tweaked between 128 to 256 time points to establish a baseline that
was manually checked for accuracy.
45
O4(x,y) = rolling ball baseline filtering(O3(x,y))
The APD map was generated from O4. The measurement was taken from the beginning
of the upstroke (moving away from the baseline) to the first occurrence of the data point
less than half of the peak value.
3.4.4 Intracellular Recordings of Cell Sheets
Micropipettes were pulled from borosilicate glass (thin walled pipettes measuring 1.2mm
outer diameter, 0.75 mm internal diameter, and thick walled pipettes measuring 1.2 mm
outer diameter and 0.4 mm internal diameter, WPI, Sarasota, FL, U.S.A) using a
Flaming/Brown pipette puller (model p-87, Sutter Instrument Company, Novato, CA,
U.S.A). Pipettes were filled with 3M KCl solution and the resistance was ~30 MΩ for
thin walled pipettes and 80-100 MΩ for thick walled pipettes. High resistance
microelectrodes were required so as to minimize cellular damage, and minimize K+
leakage. Pipettes were pulled with wispy, long shanks that allowed for flexibility of the
pipette when impaled in thin tissue. A vacuum chamber was used after filling to remove
bubbles. Pipettes were positioned using course manipulators (Newport, Irvine, CA,
U.S.A) and recordings were acquired at 1000 Hz using an Axopatch 200B amplifier with
a MiniDigi 1A digitizer and Axoscope acquisition software (Molecular Devices,
Sunnyvale, CA, U.S.A). Impalement of tissue was performed by first positioning the
microelectrode on the surface of the tissue which was evident by a drop in the potential
recording.
46
Chapter 4
Results
4.1 Generation of cardiomyocytes from hESCs
We employed the HES3 line of hESCs which had sequences encoding enhanced GFP
(eGFP) introduced into the NKX2-5 locus by homologous recombination (NKX2-5egfp/w)
which has previously demonstrated its utility and reliability in faithfully reporting
endogenous Nkx2.5 expression 208. The homeobox gene Nkx2-5 is the earliest known
marker of vertebrate heart development, expressed in early cardiac progenitor cells and
through adulthood209. This line was chosen to facilitate quantification of cardiac
differentiation and purification of cardiomyocytes when performing single cell analyses.
This human ES cell line expressed cell surface makers that characterize undifferentiated
human cells including stage-specific antigen (SSEA)-3, SSEA-4, TRA-I-60, TRA1-81.
These cells did not stain strongly for SSEA-1 (s 1) 126.
Figure 1: Flow cytometric analyses of the frequencies of cell surface markers used routinely to identify undifferentiated hESCs; stage specific embryonic antigen 3 and 4 (SSEA3, SSEA4) and two human EC cell antigens Tra-1-60 and Tra-181 also used to mark undifferentiated hESCs. SSEA1 is expressed on differentiated hESCs. Flow cytometry performed after 5 days of hESC culture on MEFs at the time point when cells are dissociated for the T0 of differentiation. Here we show, based on gates set on the unstained sample on the far left, high frequencies of makers of undifferentiated hESCs and a low frequency of a marker of differentiated hESCs.
47
To generate cardiomyocytes from hESCs, we employed the embryoid body staged
differentiation protocol that involved the formation of a primitive streak like population
defined by T (BRACHYURY) expression (days2-4), the induction and specification of
cardiac mesoderm (MESP1; days 3 and 4), and the expansion of cardiovascular lineages
154. This protocol uses the combination of Activin and BMP at the primitive streak and
mesoderm stage, and then the addition of the WNT inhibitor IWP2 as well as retinoic
acid (RA) to the developing cardiomyocytes (Figure 2). As previously mentioned, RA
signaling is crucial for atrial chamber development in vivo, and its activation has been
shown to effectively drive atrial cardiomyocyte differentiation in mouse and human ESCs
161 158 210 160.
Figure 2: Scheme of the protocol used to differentiate hESCs towards the cardiomyocyte lineage highlighting the three main stages of development: 1) mesoderm induction, 2) cardiovascular specification and 3) maturation. Protocols for the generation of “ventricular” and “atrial” cardiomyocytes differed in terms of the concentration of Activin and BMP4 used (high vs. low respectively) and whether or not retinoic acid (RA) was added to the cultures at T3. RA is added to the low Activin and BMP4 induction protocol in order to generate “atrial” cardiomyocytes as discussed in the text.
Our group has gone on to show that the manipulation of the BMP and Activin signaling
pathways can generate two distinct cardiac precursors as defined by their expression of
48
either Raldh2 (“atrial” precursosrs), or upregulation of Cyp26a1 (“ventricular”
progenitors) in a temporal pattern consistent with developmental cues (Figure 3). The
first protocol uses lower levels of BMP4 and Activin during mesoderm induction
(3B/2A), while the second uses higher concentrations (10B/6A) as discussed in the
Methods section. The 3B/2A protocol displayed a higher relative expression of RALDH2,
the main Raldh involved in early cardiac development. The 10B/6A protocol produced a
population of cells that upregulated , Cyp26a1, the cytochrome p450 enzyme that
degrades RA.
Figure 3: qRT-PCR results comparing the relative expression of ALDH1A2 and CYP26A1 over time in a protocol incorporating low levels of BMP4 and Activin (3B/2A) and a protocol with higher BMP4 and Activin (10B/6A). This highlights the different cell types generated through the manipulation of the strength of signaling in these pathways. While both protocols were previously demonstrated to make cardiomyocytes at high frequencies, the 3B/2A protocol generated a population of cells with significantly higher expression of ALDH1A2 (the gene encoding RALDH2) than the 10B/6A protocol. In parallel, the 10B/6A protocol generated a population of cells with significantly higher CYP26A1 (the cytochrome p450 enzyme that degrades RA). These graphs highlight the importance of the careful manipulation of the BMP4 and Activin pathways in directing differentiation of a cell population poised to generate and employ RA (3B/2A), in contrast to a cell population that upregulates an enzyme to protect itself from RA exposure (10B/6A). Unpublished data reproduced with permission from Dr. Stephanie Protze and Jeehoon Lee. To validate the function of the two systems, our group employed a commercially
available assay known as Aldefluor 211. Briefly, cells are incubated with a fluorescent
aldefluor substrate that when oxidized by an aldehyde dehydrogenase can no longer
49
efflux from a cell and becomes effectively trapped. The amount of fluorescence exhibited
reflects the activity of aldehyde dehydrogenase in the cell. The addition of 4-
diethylaminobenzaldehyde (DEAB), which is a potent inhibitor of Aldh enzymes,
prevents the enzymatic oxidation of this product thereby providing a negative control for
fluorescence. Our group demonstrated that the activity of aldehyde dehydrogenases was
restricted to a population of cardiac precursors exposed to lower concentrations of BMP4
and Activin, suggesting that this population of cells not only expresses Raldh2, but that
the enzyme is functional in these cells and not in the “ventricular” precursors (Figure 4).
Figure 4: Flow cytometric analyses plotting the frequencies of aldefluor positive cells against PDGFRα positivity (a marker of cardiac progenitor cells). Using the aldefluor assay, cells that express high levels of aldehyde dehydrogenase will fluoresce, in this study used as a potential functional marker of RALDH2 activity. DEAB, a potent inhibitor of Aldh enzymes, serves as a negative control. Once again, two protocols are compared incorporating either low levels of BMP4 and Activin (3B/2A) or higher BMP4 and Activin (10B/6A). While both protocols generate a high frequency of PDGFRα+ cells (an early marker of cardiomyocyte differentiation efficiency), only the 3B/2A protocol generates a high frequency of aldefluor+ cells. This functional assay reiterates the importance of precise manipulation of the BMP4 and Activin signaling pathways in patternings cells that have the machinery necessary to generate and thus utilize RA. Unpublished data reproduced with the permission of Jeehoon Lee and Dr. Stephanie Protze.
50
In order to monitor and optimize the efficiency of the differentiation protocol, several
surface makers were monitored at key commitment stages. The induction of cardiac
mesoderm was monitored by the temporal expression of CD56, KDR (Flk-1) as well as
PdgfR-α. CD56/NCAM marks the differentiation of early mesodermal progenitors from
hESCs. These mesodermal progenitors however are multipotent, giving rise to many
mesodermal lineages including cardiomyocytes 212. KDR/Flk-1 has demonstrated its
utility in marking the induction of cardiac mesoderm in mouse and human ESCs,
however it is also expressed in different mesoderm populations 142. PdgfR-α is co-
expressed with Flk-1 in the cardiac mesoderm at embryonic stages and is found on
cardiac progenitor cells in the cardiac crescent 213 143. Optimizing the induction steps to
generate greater than 60% Flk-1+ PdgfR-α + cells generates cultures of highly enriched
cardiomyocytes 155. The earliest time point at which we monitor mesodermal induction is
at day 3 or day 4, looking for the emergence of a double-positive population (>80%) of
CD56+ and Pdgfr-α+ cells. The generation of cardiomyocytes was quantified using flow
cytometry by staining of cardiac troponin T (cTNT) and the surface marker SIRPA,
which has previously been shown to uniquely mark the cardiomyocyte lineage in hPSC
differentiation cultures (Figure 5) 214.
51
Figure 5: Flow cytometric analyses plotting the frequencies of several important markers used in the optimization and validation of the differential protocols employed. The plot on the left demonstrates an optimal profile on day 4 with a high frequency of CD56+ and PdfR-α+ cells. The middle plot is representative of a good differentiation protocol which generates a high frequency of SIRPα+ cells. The frequency of CD90+ cells is also routinely assessed, and employed to optimize purification of cardiomyocytes, to quantify the frequency of CD90+ cells in the cell sheet culture, or to sort out CD90+ cells and then incorporate them into the cell sheet at controlled frequencies. The final plot on the right demonstrates an efficient differentiation protocol generating 91.9% cTNT+ cells.
4.2 Molecular markers of “atrial” and “ventricular” cardiomyocytes
Differentiation cultures were analyzed for atrial and ventricular specific markers with and
without the addition of RA using qRT-PCR (Figure 6). Cultures that were exposed to RA
compared to cultures that were not exposed to RA were enriched in the atrial specific
markers ANF (148.3 ± 9.4 vs. 16.4 ± 3.9, p < 0.0001, n = 4), KCNJ3 (0.58 ± 0.16 vs.
0.046 ± 0.01, p = 0.01, n = 4), Cx-40 (1.6 ± 0.4 vs. 0.019 ± 0.01, p <0.01, n = 5), and
CaCNA1d (13.0 ± 1.5 vs. 5.6 ± 1.0, p <0.01, n=5). Cultures that were exposed to RA
expressed significantly less ventricular specific markers compared to cultures that were
not exposed to RA including MLC2V (0.63 ± 0.3, p <0.01, n =5), and IRX4 (0.037 ±
0.02 vs. 0.46 ± 0.1, p <0.01, n =5).
52
Figure 6: qRT-PCR-based expression analyses of a control protocol (10B/6A) generating “ventricular” cardiomyocytes and the RA protocol (3B/2A+RA) incorporating lower levels of BMP4 and Activin as well as RA. Cell populations were analyzed at T20 of the differentiation protocol. The RA protocol was able to generate cells that are enriched in atrial specific markers (NPPA, KCNJ3, and GJA5) and lack ventricular markers (MYL, IRX4). Values are shown relative the housekeeping gene TBP. Error bars represent standard deviation of the mean from the values of independent experiments (N≥4); *P≤0.05, **P≤0.01, ***P≤0.001 as analyzed by Student's T-test. Unpublished data reproduced with the permission of Jeehoon Lee and Dr. Stephanie Protze.
Immunostaining confirmed the qRT-PCR findings and demonstrated that RA treated cells
expressed cTNT, but did not express MLC2v, whereas control cells expressed both
MLC2v and cTNT (Figure 7).
53
Figure 7: Fluorescent immunostaining for the presence of MLC2v and cTNT in day 20 populations. DAPI staining shows cell nuclei. On the left is the control (10B/6A) protocol generating cells that express both cTNT and the ventricular specific marker MLC2v. On the right, the atrial protocol (3B/2A+RA), that includes the addition of retinoic acid, generating cells that express cTNT but do not express MLC2v. Unpublished data reproduced with the permission of Jeehoon Lee and Dr. Stephanie Protze.
4.3 Single Cell Electrophysiology
Cardiomyocytes derived from hESCs that received RA during their differentiation
protocol (“atrial” cardiomyocytes) had predominantly atrial action potential (AP)
morphologies, whereas cardiomyocytes that were not exposed to RA (“ventricular” or
control cardiomyocytes) had predominantly ventricular action potential morphologies
(Figure 8). The “atrial” differentiation protocol, that included the addition of RA,
generated 90% atrial like APs, 5% nodal like APs and 5% ventricular like APs (n = 20
cells) (Figure 9). The control, or “ventricular” differentiation protocol, that did not
include the addition of RA, generated 85% ventricular like APs, 10% atrial like APs and
5% nodal like APs (n = 20 cells) (Figure 10). The electrophysiologic characteristics of
the recorded APs can be seen in Table 1.
54
Figure 8: Single cell patch recordings demonstrating typical APs of RA (3B/2A+RA) compare to the control protocol (10B/6A without RA). Cells are carried through to day 20, to allow for maturation and the genesis of contracting cardiomyocytes, where they were then plated as single cells and patched in Tyrode’s solution to record single cardiomyocyte APs. RA treated cells demonstrate atrial AP morphologies whereas control cells that were not exposed to RA demonstrate ventricular AP morphologies.
Figure 9: Showing the presence of cardiomyocyte types with representative AP profiles and their prevalence (n = 20 cells) in the atrial directed differentiation protocol (3B/2A + RA) used to generate atrial cardiomyocytes from hESCs. Cells are carried through to day 20, to allow for maturation and the genesis of contracting cardiomyocytes, where they were then plated as single cells and patched in Tyrode’s solution to record single CM APs. The majority of APs carried an atrial phenotype (90%).
55
Figure 10: Showing the presence of cardiomyocyte types with representative AP profiles and their prevalence (n = 20 cells) in the standard directed differentiation protocol (control, 10B/6A) used to generate cardiomyocytes from hESCs. Cells are carried through to day 20, to allow for maturation and the genesis of contracting cardiomyocytes, where they were then plated as single cells and patched in Tyrode’s solution to record single CM APs. The majority of APs carried a ventricular phenotype (85%).
N (cells)
dv/dtmax
(V/s)
DMP (mV)
APA (mV) APD50
(ms) APD90
(ms)
Control “Ventricular”
20 54 ± 3 -53 ± 1 101 ± 2 479 ± 59 547.2 ± 59
RA treated “Atrial”
20 44 ± 5 -50 ± 1 78 ± 3** 33 ± 6** 187 ± 13**
Table 1: Characteristics of APs recorded from the protocol generating atrial cardiomyocytes (3B/2A + RA) compared to the protocol used to generate primarily ventricular cardiomyocytes (10B/6A). The majority of cells demonstrated features of electrophysiologic immaturity with more positive membrane potentials, and lower upstroke velocities and action potential amplitudes. APA, action potential amplitude; APD50, action potential duration at 50% of repolarization; APD90, action potential duration at 90% of repolarization; DMP, diastolic membrane potential; dv/dtmax, maximum action potential upstroke velocity; N, cell number; **P≤0.0001 as analyzed by Student's T-test.
Comparing the cardiomyocytes generated from the “atrial” differentiation protocol to the
cardiomyocytes generated from the “ventricular” or control protocol, as would be
expected, the “atrial” cardiomyocytes had a shorter APD50 (33 ± 6 ms vs. 479 ± 59, p <
0.0001) (Figure 11) There was no significant difference between differentiation protocols
with respect to the maximum upstroke velocity, or the maximum diastolic potential. Of
56
note, both of theses values are considerably more positive than that expected of adult
tissue. In addition, “atrial” cells had a lower action potential amplitude (78 ± 3 mV vs.
101 ± 2 mV, p <0.0001), and a dramatically shorter APD90 (187 ± 13 ms vs. 547 ± 59
ms, p <0.0001).
Figure 11: Comparing the AP characteristics of cells generated from the “Ventricular” (10B/6A) differentiation protocol compared to the “Atrial” (3B/2A + RA) differentiation protocol. n = 60 APs representing 20 cells (3APs/cell). Error bars represent 95% standard deviation from the mean. **P≤0.0001 as analyzed by Student's T-test.
There was a significant correlation between the action potential amplitude and the
maximum diastolic potential (R2 = 0.373, p<0.001), which was also seen when
considering the maximum upstroke velocity and the maximum diastolic potential (R2 =
0.176, p = 0.001). Using a technique called an anode break, negative current was injected
into the cells in order to achieve a more negative resting membrane potential. This was
done repeatedly until a resting membrane potential of -70 mV was achieved, and then the
current was extinguished allowing the cell to fire a spontaneous action potential. Using
this technique, a higher action potential amplitude and maximum upstroke velocity were
achieved (Figure 12). Finally, “atrial” cardiomyocytes had a much faster spontaneous
beating rate, as demonstrated by a shorter cycle length, compared to the “ventricular”
cells (680 ± 5 ms vs. 1670 ± 12 ms, p < 0.0001).
57
Figure 12: Patch clamp recording in Tyrode’s solution of a cardiomyocyte generated from the atrial protocol (3B/2A + RA) during an anode break. The injection of negative current into the cell drove the resting membrane potential to a more negative value and then released. The subsequent action potential amplitude is clearly higher than the subsequent APs starting from a less negative resting membrane potential in the image on the left. On the right, the same cell with multiple time sequences overlayed to demonstrate the reproducibility of this maneuver in achieving higher action potential amplitudes.
4.4 MEA Electrophysiology
“Atrial” and “ventricular” cell sheets were plated on MEA chips as described. External
field potentials were successfully recorded from the MEAs of both atrial and ventricular
cell sheets (Figure 13). These are akin to an electrocardiogram measurement on patients
and have clearly demonstrable depolarization and repolarization phases facilitating
measurements of field potential durations (FPDs) which are directly linked to the action
potential duration. As expected from the single cell analyses, the FPDs of “atrial” cell
sheetss were much shorter than that of “ventricular” cell sheets. Pacing was performed
through the Cardio2D system in a unipolar configuration. Capture was reliably
demonstrated by direct visualization of the cell sheet, the demonstration of capture at the
tail end of the pacing artefact, and through routine electrophysiologic principles including
58
the perturbation of cycle length as demonstrated by the pacing cycle length, the
spontaneous beating rate, and the post-pacing interval (Figure 14).
Figure 13: MEA recording of a “ventricular” cell sheet on the left generated from the 10B/6A differentiation protocol compared to an “atrial” cell sheet on the right generated from the 3B/2A + RA protocol. Cell sheets were allowed to beat spontaneously. FPDs recorded and labeled for both cell sheets. The “atrial” cell sheet displays a shorter FPD as would be predicted by the single cell data in which atrial cells had significantly faster repolarization times. FPD = field potential duration.
Figure 14: MEA recording of “atrial” cell sheet (generated from the 3B/2A + RA differentiation protocol). Capture is demonstrated at the pacing rate and thus pacing has successfully overdriven (either suppressed in the case of automatic activity or accelerated in the setting of triggered activity) the spontaneous pacemaker. The post pacing cycle length (CL), commonly referred to as the post pacing interval, is the time required for the spontaneous activity to resume, which then returns to its spontaneous rate. The clear demonstration of these three intervals, in addition to the capture signal distinct from the pacing artefact, in this tracing strongly supports our ability to pace and capture the cell sheets at this pacing rate. Fibrillatory activity, appreciated by direct visualization of the cell sheets, was achieved in
“atrial” (n=3), but not “ventricular” (n=3) cell sheets despite aggressive burst pacing
59
protocols down to the shortest cycle length of 50 ms. Similarly, 2 point stimulation using
the S1-S2 protocol previously described was not able to induce fibrillatory activity in the
ventricular cell sheets despite numerous and aggressive pacing protocols and field
generating configurations (data not shown). Unfortunately, fibrillatory activity in the
atrial cell sheets reduced the signal to noise ratio on the recorded field potentials, and
signals were no longer distinguishable. The inability to monitor and thus map AF on the
MEA system available led us to abandon this methodology in favour of optical mapping.
4.5 Optical mapping electrophysiology
The optical mapping configuration was optimized such that activation wavefronts could
be visualized with minimal processing. Advanced processing techniques were utilized for
noise reduction and signal augmentation, but were not used for analyses. Fifty percent of
“atrial” cell sheets had a propensity to develop spontaneous fibrillatory activity within the
first week of plating, with the prevalence of fibrillatory activity decreasing over time. At
the time of optical mapping, 1 in 4 cell sheets would continue to exhibit fibrillatory
activity. This occurred for the most part in cell sheets that had areas of non-confluence,
where “holes” in the cell sheets acted as anchors for re-entry (not shown). Non-confluent
cell sheets were discarded.
In confluent cell sheets, spontaneous fibrillatory activity at the time of optical mapping
was infrequent. “Atrial” cell sheets demonstrated spontaneous beating rates of 78 ± 14
bpm, with the majority developing spontaneous activity typically at the edges of the cell
sheets where a favourable source-to-sink relationship exists215 for cardiomyocyte clusters
with the fastest intrinsic firing rate (Figure 15). Once optimized, optical action potentials
(OAPs) could be recorded in a majority of cell sheets (Figure 16).
60
Figure 15: A series of sequential still images taken from the optical recording of an atrial cell sheet (3B/2A + RA) of 1 cm diameter using 10 μM Di-4-ANEPPS and 10 μM blebbistatin and captured on an EMCCD. Electrical activity appears to originate from the top right corner of the cell sheet and propagate more rapidly along the edge of the cell sheet than towards its centre. This generates an activation wave front that ultimately circumnavigates the cell sheet and then moves relatively uniformly towards its centre, where the greatest sink to source mismatch occurs. Ultimately the wavefront extinguishes at the centre of the cell sheet, and a new wavefront is generated from the same origin and propagates in the same manner.
Figure 16: Typical optically recorded APs of an “atrial” cell sheet (3B/2A + RA) in “AF” using 10 μM Di-4-ANEPPS and 10 μM blebbistatin and captured on an EMCCD. The OAPs carry the typical morphology of atrial APs. The amplitude relates to the amount of fluorescent change of the voltage sensitive dye that occurs during a depolarization and repolarization event. The bipolar electrodes were able to pace and generate point stimulation which propagated
uniformly through the cell sheet. Increase of the pacing rate resulted in shortening of the
“atrial” APD (Figure 17). Burst on and off pacing at a cycle length of 50 ms for 1-2
minutes consistently generated “AF” as demonstrated by an increase in beating rate, the
development of continuous electrical activity in the cell sheet, and the generation of
61
rotors (Figure 18). On one occasion, electrical alternans was recorded during rapid pacing
(cycle length of 50 ms), and was followed by rotor initiation (Figure 17).
Figure 17: On the left, optical action potential duration (OAPD) measurements on the Y axis plotted as a function of rate on the X axis demonstrating the APD restitution curve of a typical “atrial” cell sheet. At faster rates, the APD restitution curve becomes increasingly steep. On the right, electrical alternans is induced in the same cell sheet at a pacing rate of 1200 bpm demonstrating cyclic, beat-to-beat variations in AP amplitude and action potential duration at a constant stimulation frequency. The cell sheet cannot be captured 1:1 at this pacing rate. After this recording, “AF” was recorded in the cell sheet upon cessation of pacing.
Figure 18: A series of sequential still images taken from the optical recording of an “atrial” cell sheet (3B/2A + RA) of 1 cm diameter using 10 μM Di-4-ANEPPS and 10 μM blebbistatin and captured on an EMCCD. A rotor continuously generates an electrical wavefront, anchored by the phase singularity. Once initiated, the rotors were extremely stable, persisting in culture for as long as 8
weeks. When comparing the APD maps of a cell sheet in sinus rhythm compared to a cell
sheet which had been induced into rotor formation and remained persistently so
chronically (for 6 weeks), there is a clear difference in the APD heterogeneity (Figure
62
19). This heterogeneity is one manifestation of the consequences of electrical remodelling
taking place in the cell sheet after rotor induction that can be identified both acutely and
chronically in this model.
Figure 19: Action potential duration (APD) maps generated from the “atrial” cell sheets using DI4-ANEPPS and an EMCCD. The figure legend represents APDs in msec. The image on the left was generated from a cell sheet prior to arrhythmia induction, while the APD map on the right represents a stable rotor that was induced with burst pacing. Note the increased APD heterogeneity in the APD map of the rotor compared to the APD map on the left. This heterogeneity is one manifestation of the nearly immediate electrical remodelling that is taking place in the cell sheet that can be identified both acutely and chronically in our system. As discussed previously, the inability of the excitation wavefront to depolarize the PS of
an ongoing rotor, and the tendency of the wavefront to curve around the PS create the
basis for rotor initiation and maintenance216. It has been previously demonstrated that the
wavefront curvature is highest around the tip resulting in slowing of the conduction
velocity near the PS, or centre of the rotor. Further from the centre, the curvature is
reduced and conduction velocity increases. Several mapping techniques have evolved to
assist in the understanding and analysis of rotor dynamics. Phase mapping was first able
to identify the PS by giving it an arbitrary phase, and its surrounding element a
continuous progression of phases equal to ±2π 99. More recently, the Hillbert transform
63
has facilitated the computing of the instantaneous phase 217. Another important
developments was the analysis of the time-dependent behaviour of rotors using a Fast
Fourier Transform Analysis and generating a frequency map of the signals of interest.
Selecting the maximum frequency in the Fourier spectrum allowed for the construction of
dominant frequency maps 218.
Cardiac sheets, predominantly using rat neonatal rat and embryonic chick ventricular
cardiomyocytes, have demonstrated their utility in generating and mapping 2-D rotors
and re-entry 219 220. In these cases, owing to the stability of rotors in this model,
activation maps were able to capture the same data content generated by phase maps and
DF maps, required of more complex and three dimensional structures. We successfully
generated activation maps (Figure 20) and conduction velocity maps (Figure 21) at both
baseline, when a single nodal source acts as a pacemaker (henceforth referred to as “sinus
rhythm” or “SR”) and after rotor induction (henceforth referred to as “AF”). Activation
maps appropriately identified the PS where depolarization and repolarization wavefronts
collided, and captured the property of increasing curvature as the excitation wavefront
propagated away from the PS. Conduction velocity maps were able to demonstrate the
inverse relationship between conduction velocity and distance from the PS.
64
Figure 20: Activation maps generated from recordings of “atrial” cell sheets using DI-4ANEPPS and an EMCCD. Optical APs are obtained on each pixel of the EMCCD and the accrued points are assigned to an isochronal color scale based on their respective activation times. On the left, a cell sheet before the induction of an arrhythmia. The cell sheet spontaneously generated a concentric ring of triggered activity which starts on the periphery of the culture and propagates uniformly to the centre (see Figure 15). This is likely a result of the aforementioned source sink relationship which generates edge effects in the culture. On the right, an activation map recorded after the induction of a rotor using burst pacing (see Figure 18). Note the increase in frequency after the induction of continuous rotor activity in the dish simulating the induction of a major driver of clinical AF. Isochronal lines are recorded in ms and labeled on the map.
Figure 21: Conduction Velocity (CV) maps generated from the activation maps in Figure 20. Again, the map on the left demonstrating baseline activity before rotor induction with uniform progression of electrical signal from the periphery of the dish to the centre. The arrows represent the vector of the electrical propagation at each point, and the size of the arrows represents the magnitude of the velocity. On the right, a CV map generated after rotor induction. Note the smaller magnitude of the arrows at the core of the rotor. This facilitates rotor stability and in fact anchors the rotor around an area of a phase singularity. Here, the convex curvature reaches a critical value preventing electrical activity to invade the core and thus maintaining the unexcitable obstacle around which the re-entrant rhythm circulates.
65
4.5.1 Optical mapping the effects of AADs
To test the appropriateness and applicability of our model system to the tasks of drug
screening and drug discovery, we first sought out to demonstrate the known effects of
commonly used AADs. Indeed in our preliminary studies we observed predictable
effects of flecainide and dofetilide on the OAP morphology (Figure 22). The major and
important limitation to these studies was the rapid photobleaching of voltage sensitive
dyes thus limiting the exposure time and therefore the dose range of each drug trialed.
Figure 22: Optical action potentials (OAPs) from cell sheets of “atrial” cardiomyocytes derived from hESCs are recorded using the voltage sensitive dye DI4-ANEPPS on an EMCCD. Representative OAPs from an “atrial” cell sheet after induction of a rotor using burst pacing are demonstrated in Figure 5A. In Figure 5B, the baseline recording (in black) is compared to a second recording after the addition of 1 μM dofetilide (in blue), a specific IKr blocker. There is prolongation of the OAP as predicted by the effects of dofetilide on the major repolarizing K+ current in these cells. Figure 5C again demonstrates another “atrial” cell sheet in which a rotor is induced and OAPs are
recorded (in black), and a second OAP recorded after the addition of 10 µM flecainide (in
red). Slowing of the upstroke of the cardiac action potential can be appreciated, as would be predicted by the sodium channel blocking properties of the drug. Interestingly, there is also OAP prolongation, which may be related to the additional K+ blocking effects of flecainide that are predicted to extend the atrial wavelength and induce rotor termination. Dofetilide is a class III anti-arrhythmic which selectively inhibits the rapid component of
the time-dependent outward potassium current (Ikr) thus causing APD prolongation
without influencing the rate of depolarization or conduction velocity 75. As predicted
however, dofetilide is not specific to the atria and thus causes APD prolongation and
66
resultant QT prolongation in the ventricles, predisposing patient to torsades-de-pointes
and sudden death. Cardiomyocytes derived from hESCs and hiPSCs have demonstrated
exquisite sensitivity to Ikr blockade exceeding that of animal tissue assays 221. Atrial
cardiomyocytes derived from hPSCs have not yet been tested for their response to Ikr
blockade, and so we chose to test dofetilide as our first AAD screen on our model.
Importantly, we chose to apply AADs after induction of “AF” to generate a clinically
relevant tissue model that would have greater translational meaning and potential.
“AF” was induced in rotors during “SR” with burst pacing as described. The induction of
“AF” in these cultures did not have an effect on the mean OAPDs across the cell sheet
(Figure 23). As would be predicted, dofetilide at a concentration of 1 μM dramatically
prolonged the OAPD of “atrial” cardiomyocytes during “AF”. When comparing cell
sheets during “AF” and then after exposure of dofetilide, there was a significant
prolongation of the OAPD (529 ± 137 vs. 809 ± 432, p = 0.02) (n =3) without a
significant effect on cycle length (687 ± 170 ms vs. 809 ± 432 ms, ns)(n=3).
Figure 23: Plotting the sequential effects of rotor induction and 1 μM dofetilide on the optically mapped APDs and cycle length of “atrial” cell sheets. The induction of a rotor decreased the cycle length (increased the rate) without a significant effect on the APD.
67
Dofetilide significantly prolonged the APD without a significant effect on cycle length. Error bars represent 95% standard deviation generated from three independent experiments (N=3); *P≤0.05 as analyzed by Student's t-test. Dofetilide alone was not able to convert “AF” to “SR” acutely (within 4 hours). The
introduction of overdrive pacing (pacing at a cycle length exceeding that of the ongoing
rotor) to the cultures that were exposed to dofetilide however was able to convert “AF” to
“SR” as demonstrated by the return of a nodal pacemaker driving the rate of the system,
and the elimination of reentry and rotor dynamics (n=2). Interestingly, on one occasion,
overdrive pacing extinguished one rotor immediately transitioning to another rotor with a
different PS that was geographically distinct from the first. This second rotor was then
eliminated by overdrive pacing leading to “SR”. Dofetilide did not have an effect on
conduction velocity across the cell sheets (n = 3)(Figure 24).
Figure 24: Demonstrating the effects of rotor induction, 100 nM dofetilide and 1 μM dofetilide on conduction velocity in “atrial” cell sheets. When considering the conduction velocity across the cell sheet, dofetilide does not impact on conduction velocity. Error bars represent 95% standard deviation generated from three independent experiments (N=3). The next anti-arrhythmic trialed was flecainide, a class Ic AAD which blocks the Nav1.5
sodium channel in the heart, slowing the rapid upstroke component of the cardiac AP.
68
Flecainide has been shown to reduce excitability and slow conduction velocity.
Flecainide also has effects on the repolarizing potassium currents which are proposed to
increase wavelength thus explaining the clinical observation of decreased AF
susceptibility for patients on this drug. Addition of flecainide to cell sheets in “AF”
appeared to have an effect on the cycle length on two cell sheets trialed (840 ± 396 ms vs
1766 ± 278 ms), with less of an impact appreciable on OAPD prolongation in both cell
sheets trialed (523 ± 119 vs to 789 ± 160 ms) (Figure 25).
Figure 25: Plotting the sequential effects of 5 μM flecainide and 10 μM flecainide on the optically mapped APDs and cycle length of a single “atrial” cell sheet after the induction of a rotor. Flecainide appears to slow cycle length in a dose dependent manner without exerting a significant independent effect on action potential duration. Also as expected, the addition of flecainide appeared to decrease conduction velocity
across the cell sheets in “AF” (3.5 ± 2.4 cm/s vs 2.0 ± 1.8 cm/s)(n=2). The addition of
flecainide did not convert the cell sheets from “AF” to “SR” (n=3) (Figure 26).
69
Figure 26: Demonstrating the effects of 5 μM flecainide and 10 μM flecainide on the conduction velocity of an “atrial” cell sheet after the induction of a rotor. Flecainide appears to have a dose dependent effect on conduction velocity as seen on both cell sheets trialed. A mean and 95% standard deviation are represented for all conduction velocities recorded from the conduction velocity map of one cell sheet. To further dissect the effects of AADs on our human model of SR and AF, we analyzed
the conduction velocity maps in tertiles according to the geographic distance from the
focal source, either the nodal pacemaker in the setting of “SR” or the PS in the setting of
“AF”. This was based on our observation that our model had recapitulated the principles
components of curvature and conduction velocity essential to rotor initiation and
maintenance, and our hypothesis that AADs may have a differential effect depending on
the location on the rotor. The induction of a rotor indeed had a dramatic effect on slowing
conduction velocity at the source of the electrical wavefront (3.8 ± 2.1 cm/s vs. 0.5 ± 0.2,
p = 0.04)(n=3) with a trend towards the inverse effect on the outer tertile of the cell sheet
70
that did not reach statistical significance. Dofetlide, as would be predicted, did not have
an effect on conduction velocity at any site along the rotor (Figure 27). Flecainide
however appeared to have a differential effect on conduction velocity slowing that was
dependent on the distance from the PS, unexpectedly slowing the CV disproportionately
at the site of the broadest curvature on the rotor (from 5.2 ± 4.9 cm/s to 2.2 ± 2.0 cm/s)
(n=2) (Figure 28).
Figure 27: Demonstrating the effect of dofetilide on conduction velocity as a function of distance from the origin of the electrical wavefront. The cell sheet was divided into thirds according to the distance from the origin. At baseline, the cell sheet has a spontaneous pacemaker which drives the cell sheet in “SR”. The origin is therefore at the site of the earliest activation in this case. After “AF” induction, the origin is defined as the phase singularity (PS), or centre of the rotor. Regions of interest are drawn over these regions of the conduction velocity maps generating means and standard deviations. There is a significant drop in the conduction velocity near the source after rotor induction (p = 0.04) which persists during the sequential uptitration of dofetilide, from 100 nM to 1 μM. Dofetilide does not appear to have any effect on conduction velocity, regardless of the distance from the PS. Error bars represent 95% standard deviation generated from three independent experiments (N=3).
71
Figure 28: Demonstrating the effect of flecainide on conduction velocity as a function of distance from the origin of the electrical wavefront. Two independent “atrial” cell sheets are displayed. The cell sheets were divided into thirds according to the distance from the origin. Displaying only data obtained after the induction of a rotor with the centre of the rotor defined as the phase singularity (PS). Regions of interest are drawn over these regions of the conduction velocity maps generating means and standard deviations. After rotor induction, as before, the conduction velocity is slowest at the centre of the rotor and increases with increasing distance from the core. The addition of flecainide appears to have a dose dependent effect on slowing conduction velocity at the middle and outer third of the cell sheet, yet does not appear to exert any influence on the rotor’s core.
72
Chapter 5
Discussion
Atrial fibrillation is an important contributor to the morbidity and mortality of Canadians,
and is a major and growing burden on our health care system and our economy1-9. While
the pace of discovery and innovation in the understanding and treatment of AF has
accelerated rapidly over the last two decades, the initial description of the disease nearly a
century ago has withstood the test of time and remained remarkably accurate despite the
relatively primitive tools of the age. A series of critical reviews in 1924 outlined the
clinical principles of AF, and the general hypotheses regarding AF initiation and
maintenance that continue to be hotly debated today 222 223. Perhaps stemming from this
confusion regarding the mechanistic basis of the disease, clinicians have limited tools
with which to target and treat the substrate that drives recurrence and facilitates the
remodeling process involved in the disease’s propensity for self-promotion. Rotor theory
has become a pre-eminent component of AF mechanistic study and treatment owing to
technologic innovations that have facilitated a greater appreciation of its importance, and
a more detailed understanding of its components. Over a parallel, yet only recently
overlapping time course, there has been a rapid advancement of hPSC technology in the
generation of pure lineages of well characterized cardiomyocytes in vitro through the
application of the fundamentals of developmental biology.
A major step forward in the application of hPSC derived cardiomyocytes to disease
modeling has been the careful dissection of embryonic cues which drive cell specific
fates in the embryo, and mimicking them in vitro to differentiate hPSCs into pure
lineages of cardiomyocyte subpopulations. While the ability of hPSC derived
73
cardiomyocytes to model electrical cardiac disorders at a single cell level, predominantly
in ventricular like cells, has been exploited in a number of disease phenotypes and drug
screens, to our knowledge this is the first study to examine atrial tissue in the setting of a
tissue based arrhythmia. This has important implications for the study of AF and the
search for novel therapies, as well as significant potential to translate our growing
understanding of the genetic contributors of AF directly to the application of personalized
medicine.
5.1 Molecular markers of “atrial” and “ventricular” cardiomyocytes
Our group has developed a differentiation protocol that has generated an enriched
population of atrial like cells through the manipulation of BMP and Activin signaling,
and the addition of retinoic acid. The cells have undergone thorough molecular
characterization which has demonstrated their similarities to atrial cardiomyocytes and
their distinct signature compared to adult ventricular cardiomyocytes, and ventricular like
cardiomyocytes generated from hESCs. Specifically, the atrial like cell population is
enriched with the atrial markers ANF, KCNJ3, Cx40 and Cav1.3, whereas the ventricular
like cell population is enriched in MLC-2v and IRX4. Functional analyses of these cells
through patch clamping has demonstrated their clear and distinct electrophysiologic
properties which again mimic atrial cardiomyocytes.
5.2 Single cell electrophysiology
We have shown that the atrial like cell population consists of an overwhelming majority
of cells displaying an atrial AP morphology (90%), whereas the ventricular like cell
population equally exhibits a majority of cells that display a ventricular AP morphology
(85%). These cell populations importantly differ in their APD50 profiles, with atrial like
74
cells displaying dramatically shorter APD50s. As previously mentioned, the single cell
electrophysiologic properties recorded suggest that these cells are immature compared to
their adult counterparts. This finding is not unexpected given the fact that there are
known differences between adult and fetal cardiomyocytes in terms of their complement
of ion channels and the resultant AP profile224.
Adult ventricular and atrial cardiomyocytes exhibit a resting phase (phase 4) in which the
resting membrane potential (RMP) does not change. This is primarily due to the
rectifying potassium current (Ik1) which stabilized the RMP at ~-85mV, which is the
reversal potential of K+. In hPSC derived cardiomyocytes, the RMP is much less
negative, a property which has been attributed to the lower expression of Ik1156,225. We,
and others, have indirectly shown that the relatively depolarized state of the hPSC
cardiomyocytes leads to a decrease in the functional availability of Na+ channels by
driving the membrane potential to more negative values and demonstrating an increase
the action potential amplitude and upstroke velocity 65,156. An additional explanation for
the relatively low upstroke velocity and action potential amplitude seen in hPSC derived
cardiomyocytes compared to published values for adult cardiomyocytes may be that cells
were patched at room temperature since the principal component of these measures is
sodium current. Sodium channels are temperature dependent, and thus cells analyzed at
37 °C may demonstrate more sodium current, and thus higher values for upstroke
velocity and action potential amplitude. Cardiomyocytes derived from hPSCs also
demonstrate large pacemaker currents (funny current, If), which is very low in adult
cardiomyocytes, causing diastolic depolarization and spontaneous contraction of hPSCs
in culture as seen in our study.
75
There are many other features of immaturity not directly addressed in our study including
morphologic parameters, sarcomeric organization, calcium handling and excitation-
contraction coupling, and metabolism224. There have been many proposed methods to
promote maturation of hPSC derived cardiomyocytes including increasing time in
culture, electrical stimulation, application of mechanical strain, chemically inducing
maturation, and the promotion of maturation by incorporating other cells types or
extracellular substrates into culture or engineered 3-D structures. In spite of all of these
efforts, many of the features of maturation have only been seen in isolation, and certain
features of the adult phenotype have never been reproduced in culture such as the present
of T-tubules.
5.3 Modeling atrial diseases using hPSCs
Despite the known limitations related to the immaturity of hPSC derived cardiomyocytes,
a fully mature state is not felt to be a necessary pre-requisite to the use of these cells to
model disease. The FDA in fact has recently proposed a directive that all new drugs
should be tested for their effects on all ion channels in human cardiomyocytes, and have
suggested that hPSC derived cardiomyocytes were an ideal model and may be
appropriate for this purpose224. While the weight of evidence to support this notion in
ventricular like hPSC derived cardiomyocytes has already been reviewed in the
background section, there has been a recent study demonstrating the utility of atrial like
cells in drug screening and drug discovery. By employing the same HES3 NKX2-5egfp/w
cell line as we have employed for our studies, Devalla et. al, generated atrial like cells
through the manipulation of BMP4, Activin and Wnt signaling, as well as the addition of
RA 226. Similar to our findings, they showed that the expression of atrial specific ion
76
channel genes was confined to their atrial like cells compared to their ventricular like
controls. Using patch clamping, they went on to show that in response to multiple atrial
specific ion channel blockers; vernakalant and Kv1.5 blocker XEN-D0101, hESC derived
atrial but not ventricular like cells demonstrated AP prolongation at the single cell level.
Additionally, and as would be expected of atrial cells, a novel Kir3.1/3.4 blocker restored
the AP shortening caused by carbachol in the hESC derived atrial cardiomyocytes. Taken
together, these results convincingly demonstrate that the cell line employed in our study,
when differentiated in cardiomyocytes using a very similar protocol to the one our group
has independently developed, generates a cell type that is appropriate, at least at the
single cell level, for the pre-clinical testing of atrial specific antiarrhythmics. We have
taken this concept a step further by demonstrating reliable and predictable drug responses
of atrial tissue in the setting of a clinically relevant arrhythmia model.
The novelty and strength of our study is the tissue based model of disease which
encompasses many aspects of clinical arrhythmology that cannot be studied at the single
cell level. In particular, we have focused on the determinants of the atrial wavelength,
conduction velocity and action potential duration, as well as the fundamental properties
of rotor dynamics. Optical mapping is a well known method to monitor and analyze the
propagation of cell excitation in whole cardiac tissues, and thus we set out to employ this
tool to study the functional properties of our multicellular network of “atrial” cells at
baseline and with the induction of complex phenomena such as reentry and spiral wave
propagation.
In our first efforts to validate our model, we have shown that “atrial” cell sheets can
demonstrate properties of electrical wavefront activation and propagation which can be
77
captured on our optical mapping setup. We have successfully designed bipolar electrodes
capable of capturing the cell sheets reliably, and generated point source stimulation with
which to control and test our cell sheets. We have gone further to demonstrate the normal
physiologic property of APD restitution in our “atrial” cell sheets, and taken advantage of
this principle to generate “AF” by driving the cell sheets to the steepest part of their APD
restitution curves to facilitate self-sustaining oscillations. We have for the first time,
observed electrical alternans at rapid pacing rates preceding the onset of “AF” in human
“atrial” tissue. Electrical alternans is a recognized clinical risk factor for cardiac
arrhythmias including AF.
“AF” in our model system was defined by spiral wave activity that carried several
signatures of rotor dynamics including the ability to self-sustain a PS, the generation of a
curved wavefront with its maximal curvature and slowest conduction velocity centered on
the PS, the ability to overdrive the system with the addition of AADs that extend
wavelength, the spontaneous initiation of a distinct PS after overdrive pacing of the
original PS, and the ability of overdrive pacing and AADs to convert the reentrant
arrhythmia. We have successfully optically mapped the induced arrhythmia to generate
activation maps, measure optical APs in the setting of an arrhythmia, and generated
vector and conduction velocity maps giving us the unique ability to measure local
conduction velocity.
78
5.3.1 APD restitution and heterogeneity
We have shown that OAPs across a cell sheet in “AF” display greater heterogeneity than
OAPs in “SR” (Figure 18). Dispersion of APDs in the atria occur under normal
physiologic conditions in humans, both in newborns and adults, and has been
documented in a number of other animal species 42,227 228. The APD and effective
refractory period shorten progressively with increased distance form the SAN, a finding
which stems from the initial observation that the atrial AP shape changes as a function of
distance and independent from the excitation sequence. This has been proposed to
protect the atria from anatomic reentry because of the resistance to unidirectional block in
the setting of an ectopic focus, and ascribed to changes in regional ionic properties 229 230.
In computer modeling studies, APD heterogeneity has indeed demonstrated a protective
effect on the development of anatomic reentry, but in contrast, did not afford protection
from functional reentry as would be seen in rotor formation 231. Wavebreak was in fact
more notable in areas of increased APD gradient leading way to stable rotor formation.
It has been proposed that such gradients contribute to the establishment of acute AF in the
structurally normal heart 232. In human AF, the atrium can achieve exceedingly high
frequencies, and thus extreme shortening of the APD would be required for APD
maintenance. Mechanisms that increased APD heterogeneity and regional abbreviation in
areas that achieve the highest frequencies, such as rotors driving AF, would lead to
increased stability of microreentrant sources by reducing interaction between the front of
the wavefront and its tail, and allowing for rapid curling of the wavefront around the PS
12 218. The APD maps generated in “SR” and “AF” indeed demonstrate a significant
difference in APD heterogeneity (ΔAPD 38 in SR vs. ΔAPD 191 in AF). At the core,
79
there is a clear boundary where the wavefront, with a very short APD, meets the wavetail
with a longer APD. It is not surprising however that we did not observe a dramatic
change in the minimum APD observed because the AF model did not induce the dramatic
increases in frequency that are observed in humans (up to 16-18 Hz), likely relating to the
conduction velocity as the rate limiting component.
5.3.2 Conduction velocity
The conduction velocities observed in our model system are much slower than that seen
in cardiac tissue in vivo. There are several possible explanations for this finding, which
has been recapitulated by many other groups using hPSCs in a cell sheet format 224. As
discussed, the RMP of the cells is less negative than that seen in vivo, leading to reduced
availability of sodium channels which are a major contributor to the propagation velocity.
This was demonstrated in our single cell patch clamping studies, and has also been
observed in micropipette recordings of our cell sheets (Appendix Figure 1).
Another major contributor to conduction velocity is the density and composition of gap
junctions. Conduction velocity however is determined by the structural organization of
gap junctions as well which co-localize with sodium channels at the intercalated disks.
Intercalated disks are found on the shortest edge of two neighbouring cardiomyocytes
thus facilitating more rapid conduction of electrical signals in the longitudinal direction
233. In cell sheets of hPSCs, gap junctions are in contrast found on all sides of the
membrane 234. A recent study has however demonstrated that over time and up to one
month in culture, hPSC derived cardiomyocytes display increasingly developed
intercalated disks, including attachment zones and gap junctions, compared to at baseline
235. This increasing degree of ultrastructural organization was associated with increasing
80
conduction velocity in the cell sheets over time. We hypothesize that this is the
underlying mechanism for our observation that spontaneous rotors were less prevalent in
our “atrial” cell sheets over time. Finally, the non-cardiomyocyte population in our cell
sheet cultures could significantly impact on conduction velocity.
Most important for the applicability of our model to serve as a drug screen is the response
of our system to AADs with known mechanisms of action. The first AAD that we chose
to test in our model system was Dofetilide, a class III anti-arrhythmic that selectively
inhibits the rapid component of the time-dependent outward potassium current (Ikr) 75.
Dofetilide is a potent AAD which has demonstrated clear utility in the management of
AF 78, however its use has been restricted by its lack of chamber specificity, causing
significant APD prolongation of the ventricles thus putting patients at risk of sudden
cardiac death. As previously mentioned, hPSC derived cardiomyocytes have previously
been shown to be more sensitive to the effects of Ikr blockade than other animal models,
implicating Ikr as a principal component of hPSC cardiomyocyte repolarization221. Our
study is the first to demonstrate the impact of dofetilide on APD prolongation in hESC
derived atrial like cardiomyocytes, the first to demonstrate OAPD prolongation in a tissue
model of hPSC derived cardiomyocytes, and the first to interrogate the effects of
dofetilide on an ongoing rotor in an hESC derived tissue model of AF. As we would
expect, dofetilide had no effect on conduction velocity over the cell sheets, and did not
disturb conduction velocity at any site of the ongoing spiral wave.
5.3.3 Effects of dofetilide on “atrial” cell sheets
Dofetilide alone did not convert “AF” in our dish to “SR”. Previous studies have
suggested that atrial wavelength plays a critical role in the antifibrillatory action of AADs
81
as wavelength prolongation was associated with atrial refractory period, and mapping of
AF on AADs showed a reduction in the number of wavelets observed 58. Other groups
have proposed that the antifibrillatory effects are indirectly associated with wavelength
prolongation, and directly linked to prolongation of the excitable gap during periods of
AF236. According to the multiple-wavelet hypothesis, the stability of AF is determined by
the number of wavelets. Widening of the excitable gap would lower the chance that
fibrillation waves encounter areas of partially refractory tissue and thus slowing, and
fractionation of wavelets will occur less frequently. Furthermore, widening of the
excitable gap would promote fusion, and on balance, the generation and extinction of
fibrillatory waves would be expected to be reduced. We have demonstrated our ability to
overdrive pace ongoing rotors easily when treated with dofetilide, which was not the case
prior to the application of the drug. This indirectly points to dofetilide’s action on
extending the excitable gap, and supports the notion that fusion and extinction of rotors
are important components of the conversion of AF to SR. Further in support of this
theory are the mechanisms of action of flecainide, the second antiarrhythmic trialed in
our model.
5.3.4 Effects of flecainide on “atrial” cell sheets
We chose to study flecainide as it is a commonly used AAD for the treatment of AF.
Flecainide is a class I antiarrhythmic agent known to depress the maximum upstroke
velocity of APs in atrial and ventricular tissue relating to a block of the sodium current
(INa) 237. Flecainide has been shown in a number of models to decrease conduction
velocity significantly in the setting of AF without a significant effect on ERP 238. During
functional reentry, the electrical wavefront is circulating around an area of functional
82
conduction block, often making sharp turns at pivot points. This would predict that the
wavefront would be forced to increase its curvature and thus set up a source-sink
mismatch resulting in conduction delay at the pivot points. This mismatch could
theoretically be preferentially aggravated by flecainide’s effects on slowing of conduction
velocity, and has been proposed as a mechanism explaining flecainide’s clinical efficacy.
Since AF relies upon the random reentry effects of wavelets entering an area previously
activated by another wavelet, drugs that increase the size of the functional circuit, and
decrease the number of wavelets available, would in effect increase the excitable gap.
The dogma that flecainide has no effect on ERP has been questioned, and may in fact be
specific to the AF model in question. Flecainide has been shown to terminate AF in the
vagotonic model of canine AF by causing tachycardia dependent increases in atrial
refractoriness and wavelength 58. This has been attributed to the drug’s effect on
decreasing atrial APD accommodation during periods of increased heart rate, possibly
related to an effect on repolarizing potassium current, as well as decreasing the
heterogeneity of atrial activation 85. Rate-dependent sodium channel block however could
also explain the observed prolongation in refractoriness independently of changes in
repolarization by depressing atrial excitability. Complicating the issue further is the
observation that flecainide has shown differential effects on potassium currents in animal
models and humans. As an example, the Kv1.5-based human Ikur has been proposed as a
potentially interesting target for novel therapies of AF, with numerous compounds now
undergoing clinical testing. The human channel is however resistant to flecainide,
whereas the dog counterpart has been shown to be quite sensitive, highlighting the
importance of using human tissue for drug screening 239.
83
We have shown that flecainide appears to have an effect on conduction velocity slowing
in our model of AF. We also noted flecainide’s effect on cycle length without a clear
independent effect of APD prolongation. While this data is clearly preliminary owing to
the limited number of repeated experiments performed, we plan on further testing to
determine the reproducibility of our findings. The following discussion is based on our
preliminary observations which will form the basis for hypotheses that have yet to be
tested.
As previously discussed, the known clinically efficacy of class Ic drugs is seemingly
paradoxic in light of our understanding of wavelength dependent reentry when
considering only the effects of conduction velocity. Anything that slows conduction
velocity without altering refractoriness would be expected to decrease the wavelength and
thus increase the propensity of tissue to develop reentrant arrhythmias. We have in fact
clearly demonstrated this principal in our model with our unique ability to optically map
rotors, and their response to drug effects. On one occasion, the addition of flecainide to a
stable rotor induced by burst pacing had a dramatic effect on conduction velocity slowing
that was not accompanied by a substantial increase in APD. This slowed the cycle
length of the rotor and facilitated the initiation of two separate rotors in an “atrial” cell
sheet (Appendix Figure 2). This finding, although predictable, would argue against the
notion that flecainide decreases the number of wavelets in AF and thus would require an
alternate explanation for its observed anti-fibrillatory effects. We propose an alternate
mechanism, based on early preliminary findings that relate to the regional variation of
drug effects along the curvature of an ongoing rotor.
84
Flecainide has been shown to produce rate-dependent reductions in atrial conduction
velocity as well as reducing the heterogeneity of wavelengths across the atria 85. We have
observed regional variation in the effects of conduction velocity slowing with the
addition of flecainide to “AF” in our model. There was an effect on conduction velocity
slowing at points along the spiral wave that had the broadest curvature, where the
conduction velocity was greatest at baseline, while there was seemingly no effect on
conduction velocity at the centre of the spiral where the wavefront curled tightly around
the PS, corresponding to the area with the slowest conduction velocity at baseline.
Flecainide is known to cause potent voltage-and frequency-dependent inhibition of
cardiac Na+ channels. Voltage gated sodium channels can exist in three distinct states;
deactivated (closed), activated (open) or inactivated (closed). With the membrane at its
resting potential, the Na+ channels are in their deactivated states. Activation gates
subsequently open in response to an electrical current, thus allowing Na+ ions to flow
into the cell and causing the membrane potential to increase and thus depolarize. At a
specific membrane potential, the Na+ channels become inactivated until the membrane
potential becomes negative enough and the cell returns to its deactivated state, ready to
participate in the subsequent depolarization event.
Flecainide has a strong preference for binding to the activated states of the sodium
channel (ie open and inactivated) and its effect is enhanced by rapid repetitive
depolarizations, and increases over the range of voltages where the channels activate 240.
This provides one potential explanation for our preliminary observation in that the centre
of the rotor may be in a relatively deactivated state, while cells along the broader
85
curvature that demonstrate a more rapid conduction velocity before treatment with
flecainide, are in relatively activated states and thus preferentially effected by flecainide.
This would certainly be supported by the observation that conduction velocity in the rotor
is lowest at its center and increases as a function of distance from the centre if the
observed conduction velocity reflects the active and inactivated states of the sodium
channel. It however directly contradicts the observations made by Allessie in a commonly
used and referenced goat model of AF. Allessie described a preferential effect of class I
drugs at areas where the rotor curvature was steepest, preferentially depressing
conduction velocity of wavelets a their pivot points 236.
An alternative explanation for our preliminary finding is that it is in fact calcium currents
and not sodium currents that predominate and drive the wavelets surrounding the rotor’s
core in our model. It has been well established that patients who have had prolonged
durations of AF show less of a response to both Na+ and K+ blocking agents 241. This has
been a proposed consequence of changes in ion channel function during atrial
remodelling in response to tachycardia and AF, a phenomenon known as “AF begets
AF”. In light of our observation, it is interesting to reflect on the fact that T-type Ca+
channel blockers appear to suppress atrial remodelling whereas L-type Ca+ channels
blockers are ineffective. In addition, Bepridil which acts on both L-type and T-type Ca+
channels has an unusual ability to convert long-standing AF242. It is possible therefore
that this hypothesis holds true in human AF. Our enthusiasm for this observation is
tempered by questions relating to the fundamental translational capacity of our model,
specifically when interrogating ion channels that have a signature in our model that is
quite distinct from the adult phenotype.
86
A feature of electrophysiologic immaturity, not previously discussed in this work, is the
expression of T-type calcium channels. These channels help to confer automaticity in
differentiating cardiomyocytes 243. In embryonic and neonatal cardiomyocytes, T-type
calcium channels have been observed at higher densities and their expression nearly
disappears in the working myocardium of the adult heart. Although difficult to quantify
in terms of current because of the lack of specific blockers, greater than 50% of hiPSC
derived cardiomyocytes display functional ICa,T244. If we are therefore able to go on to
conclusively demonstrate that the regional variation in flecainide’s effect on our model
system is dependent on the function of T-type Ca+ channels near the rotor, we would
have to go on to demonstrate the applicability of this finding to the human system. This
is of course a limitation of all model systems, emphasized here to contrast cell types,
highlight limitations, and provide a cautionary note on the translation of this novel model.
87
Chapter 6
Future Directions
The immediate path forward is clear, and our experience to date has taught us important
lessons regarding the limitations related to our current model system. Our future
directions are divided into three main themes: 1) the validation of findings to date 2) the
continued interrogation, refinement and improvement of our model system to study atrial
remodeling 3) adapting our model to answer new fundamental and clinically relevant
questions related to the identification, prevention and treatment of AF.
6.1 Validation of current findings
The principle outcomes under study have been the components of the atrial wavelength,
namely the action potential duration and the conduction velocity. In order to validate our
findings, we have incorporated micropipette recordings into our optical mapping setup.
We have successfully recorded intracellular potentials using high resistance
micropipettes, and will use these to quantify the effects of AADs on the AP morphology
and APD measurements as a standard with which to judge the accuracy of our OAP
recordings. Using this technique we have also begun to validate our conduction velocity
measurements, and generate restitution curves for conduction velocity as a function of
rate, using our stimulation probes and micropipette recordings. By measuring the time
taken from stimulation on one end of the cell sheet to the depolarization of a cell on the
opposite side of the cell sheet, and measuring the distance travelled, we have begun to
generate accurate measurements of conduction velocity while simultaneously optical
mapping to ensure point source stimulation and uniform unidirectional propagation.
Finally, we are generating “ventricular” cell sheets in order to contrast our findings with a
88
distinct cell type which should display differences in the determinants of wavelength at
baseline as well as different responses to AADs. We have already observed a significant
decrease in the propensity of “ventricular” cell sheets to generate rotors using our MEA
protocol, and expect to demonstrate the same findings in our optically mapped cell sheets.
To further validate our findings, we plan on testing AADs that have atrial specific effects
and demonstrating the predictable nature of responses in our “atrial” cell sheets compared
to “ventricular” cell sheets. This would include more recently proposed agents for the
treatment of AF: Tertiapin-Q, an IKAch blocking agent, and Ca-activated K+ channels
blockers. In addition, we propose to analyze the effects of vernakalant, which
preferentially blocks atrial-specific Kv1.5-based IKur in many models including atrial like
hESC derived cardiomyocytes. Vernakalant is used exclusively for cardioversion with
high efficacy, and thus is of interest to be tested in our model that has not yet
demonstrated purely pharmacologic cardioversion 245-247
In order to further interrogate our model, we plan on examining the structural
components of our cell sheets in greater detail. Our first step will be the identification
and quantification of Cx-40, a gap junction that is selectively expressed in atria. Another
principle component of conduction velocity and its heterogeneity across cardiac tissue
that we have not yet addressed is anisotropy, which will be interrogated by analyzing the
gross arrangement of cells in culture through immunostaining and/or electron
microscopy. Once quantified, it would be ideal to incorporate these structural measures,
as well as the ionic determinants of wave propagation, into a mathematical model that
could complement findings from the in vitro model as well as help direct future
experimentation. There are 5 models of human atrial electrophysiology that have been
89
published248. Our in vitro model could help to evaluate and refine these models.
Collectively these tools could serve to provide an in vitro and in silico assessment of
known and novel therapeutics.
The major limitation of our study has been our reliance on currently available voltage
sensitive dyes. We have trialed 3 such dyes (Di-4-ANEPPS, Di-8-ANEPPS, and RH237)
in our model system, and ultimately chose to optimize and utilize Di-4-ANEPPS as it
displayed the best signal to noise (S:N) ratio on our optical mapping setup. These dyes all
have the limitation of generating poor S:N ratios in the cell sheet format, and rapidly
photobleaching thus limiting the recording time available when using a high intensity
light source. For future studies we will use hESCs expressing Arclight249, with better (~2-
fold) S:N, and virtually no photobleaching compared to Di-4-ANNEPS249. ArcLight is a
novel genetically encoded voltage probe, developed by fusing the voltage sensor domain
of the Ciona intestinalis voltage-sensitive phosphatase to a super ecliptic pHlurion
carrying the point mutation A227D 250. Arclight has already demonstrated its utility in
measuring APs noninvasively using optical mapping of hESC derived cardiomyocytes
which quantitatively and qualitatively track with patch clamp recordings 249. Arclight is
typically introduced into cells using transient transfection or lentiviral transduction. We
have already acquired an hiPSC line that stably expresses Arclight from our collaborator.
This line has successfully been differentiated using the EB protocol previously described.
The derivative cardiomyocytes generated OAPs with excellent S:N that do not exhibit
perceptible signs of photobleaching (Appendix Figure 3). Arclight has a relatively
sluggish response time compared to Di-4-ANNEPS which will significantly affect
estimates of AP upstroke (underestimated by ~30%), however will impact immeasurably
90
on key mapping results (activation/phase, CV and APD). We will overcome this
limitation by recording APs using micropipettes, as previously described, which will be
used to "calibrate" optical data.
Extracellular field recordings (ALA multi-electrode array or MEA system) will be used
to: a) monitor electrical activity thereby (crudely) validating our activation/CV plots, b)
assess the P-wave and local electrogram “fractionation”, which links to arrhythmia
susceptibility251-253, and 3) induce and terminate arrhythmias with appropriate pacing.
This last point will be a necessary component of future studies, removing the necessity of
introducing an unsterile pacing electrode into the tissue culture medium and thus
preserving it for repeated studies. The incorporation of a stable and sterile pacing
platform as well as the utility of Arclight in repeated measures over time, will allow us to
interrogate a question that remains fundamental to the understanding and management of
AF and has been alluded to repeatedly in this document, that is “AF begets AF”.
6.2 Study AF remodeling
Regardless of its ultimate cause, AF invariably begins as isolated short-lived episodes
called "paroxysmal AF". Each paroxysm accelerates atrial remodeling thereby promoting
"persistent AF"42,43. The functional, structural, electrical and biochemical changes in atria
leading to increased AF vulnerability, providing the substrate for AF maintenance, and
making cardioversion increasingly difficult, are called remodeling14,15,254. Although some
clinical predictors of the transition from paroxysmal to persistent AF have been
identified, the molecular, biochemical and electrophysiologic mechanisms for this
remodeling remain largely unknown255,256. We propose to utilize our novel model of AF
to study the molecular, cellular and electrical changes underlying AF remodeling. As AF-
91
begets-AF, we hypothesize that rotor induction will drive time-dependent remodeling.
Indeed, APD and CV heterogeneity have been shown to develop in our cell sheets with
rotors. Moreover, we observe spontaneous rotors in a small subset of our cell sheets
which implies the presence of triggers. Thus, our cultures display dynamic changes in
wavelength (APD and CV) as well as triggers, the two key components of the AF
syndrome, which is unlike mouse/goat/dog/sheep animal models in which triggering is
externally induced.
To look at the effects of AF on remodelling in our “atrial” cardiomyocyte cultures, we
propose to make measurements at 3 time points after the induction of stable rotors: after 1
day, representing early clinical AF, after 7 days, representing the clinical time point used
to diagnose persistent AF, and after 30 days, representing chronic persistent AF. Control
cultures will be maintained in regular “sinus rhythm” (“SR”) for equivalent periods. We
will also study the effects of reverse remodelling; if possible, “SR” will be restored in the
chronic persistent “AF” group and then studied 1 week later along with a time matched
control. Since AF is ultimately an electrical phenomenon, we will use our optical
mapping rig to interrogate the electrical remodelling process as previously described. Of
course with the ability to make repeated optical mapping recordings over time, multiple
timepoints can be chosen in each cell sheet until they reach a “terminal” endpoint, such as
enzymatic digestion for single cell analysis. In addition we will be able to interrogate
other important measures of electrical remodelling that require exposure times that
exceed the capabilities of Di-4-ANEPPS because of the photobleaching and toxicity
previously described.
92
Using the Arclight line we will be able to routinely measure the effective refractory
period (ERP) with programmed external stimulation, as is routinely done in human
subjects. This is achieved using trains of 8 field pulses at 500 msec (S1) followed by
variable shorter cycle pulses (S2) until capture does not occur. S1 is also varied (400,
300, 200) allowing ERP to be determined at multiple cycle lengths. We expect that the
longer cell sheets are kept in "AF", the easier it will be to generate rotors, and these
differences will correlate systematically with differences in (inter-related) rotor/electrical
properties such as257-260: shorter APD, shorter ERP, faster rotor frequencies, greater rotor
curvatures, slower CV, and more complex activation patterns or wave breaks261. Resting
membrane potentials may become more negative262 (more background IK1). These
hypotheses will be tested by performing single cell cardiac electrophysiology studies
using patch clamping in order to dissect the observed alterations caused by electrical
remodelling into the functions and interactions of single channels and currents. To
complement these single cell studies, RNA will be isolated for targeted qRT-PCR
analysis. A future consideration could be the inclusion of RNA seq analysis.
To this point we have focused on the rapid pacing model of AF. We propose however to
test multiple acquired models of AF in the future using the aforementioned study design.
We currently are focused on three potential acquired stressors that we propose to include
in future studies; elevated parasympathetic activity, stretch, and fibrosis.
6.3 Model other types and features of acquired AF
Previous studies have established the involvement of elevated parasympathetic activity
and related IK,ACH activity in AF which has resulted in efforts to develop blockers of these
channels263. We will look for changes in the expression pattern of IK,ACH, GIRK1 and
93
GRK4, M2- muscarinic receptors, and RGS regulator proteins, as well as constitutive
activity of GRK-related channels264. We will also perform studies and analyses in the
presence of low levels of the stable muscarinic agonist, carbachol. We predict that low
levels of carbachol will be associated with electrophysiologic markers of maturation, and
that higher levels of carbachol will facilitate wavebreak and multiple wavelets
degenerating into chaotic activity.
Elevated venous/atrial pressure and atrial stretch are major factors in AF strongly
stimulating cardiac remodeling via signaling pathways leading to increased oxidative
stress 18,265-2676,265,266,267. Atrial stretch shortens atrial ERP through increased IKAch, Icl, and
Ik1268,269as well as ICa,L inactivation270 and reduced gap junctions271. We plan on
investigating the effects of cyclical stretch on hESC derived cardiomyocytes and looking
at measures of maturation and remodeling.
Finally, since many acquired forms of AF demonstrate increased interstitial atrial fibrosis,
which correlates inversely with treatment success272-275, we will combine cardiac
fibroblasts with our atrial CMs. Our initial experiments either used a cell sorting strategy
to purify the population to contain only cardiomyocytes, or incorporated ill defined
CD90+ population. We propose to repeat the same studies in cultures containing various
percentages of cardiac fibroblasts generated from hESCs using a published protocol276. It
has already been established that CD90+ cells improve maturation of cardiac gene
expression patterns, and various structural/functional measures277. In addition, fibroblasts
secrete endothelin-1 which promotes the generation and persistence of myofibroblasts,
and induce the expression of a wide variety of ECM components, including collagen type
I.
94
In contrast to other fibroblast sources, generating cardiac fibroblasts from hESCs allows
for the production of pure unlimited numbers of well characterized cardiac fibroblasts
that do not proliferate uncontrollably in our backbone culture medium. These are novel
studies with uncertain outcomes; nevertheless, we do expect that the cell sheets in which
“AF” has been induced will have increased accumulation of fibroblasts and connective
tissue since AF induces atrial fibroblasts (myofibroblasts) to proliferate and deposit
extracellular matrix278-281. This is expected to alter tissue architecture and myocyte-to-
myocyte electrical coupling as well as to create additional capacitive loads on CMs282.
We anticipate that the end result will be collective interference with electrical conduction
and increased propensity to arrhythmia initiation and maintenance. All of these properties
can be readily determined using the approaches detailed already. This novel model
system could be a powerful platform for developing new comprehensive therapies for
treating AF that specifically target fibrosis283,284.
6.4 Introduce genetic variation as a determinant and model of AF
Building on the lessons learned from our model of acquired AF, we propose to study
heritable forms of AF using hiPSCs and/or genomic editing. We have learned from the
Framingham studies that 58% of AF patients have no obvious cardiac disease or risk
factors285. AF incidence shows a strong heritability, and a genetic predisposition has been
shown to contribute to AF risk. Recently numerous AF associated mutations, candidate
genes and risk loci have been identified, yet few functional studies have deciphered their
mechanisms of action or validated their role in disease. We propose to employ our model
of human “atrial” cardiomyocytes derived from hESCs to study the underlying
mechanism for AF vulnerability in these patients with the potential to identify novel
95
therapies. As a first test of our hypothesis, we will use the CRISPR/Cas9 system to
introduce genetic changes into a background of normal hESCs. As a test case, we propose
to study the effects of GATA6 (heterozygous and homozygous) mutations in our system
286. Autosomal dominant mutations in GATA6 have been identified in cases of loneAF287.
GATA6 was chosen for several reasons; firstly, it has recently been linked to AF in rare
families with mendelian inheritance patterns and strong penetrance and expressivity.
Second, it is a transcription factor expressed in the early cardiac mesoderm 288. Third,
although GATA6 has been implicated in normal gap junction expression during periods
of stress or injury in the atria289, GATA6 is likely (by virtue of being a transcriptional
factor) to affect the expression of many genes and could provide some interesting clues
into multiple functional genes impacting on AF, which could yield greater insights into
AF mechanisms more broadly. To complement this data set, we propose to differentiate
pure populations of nodal, atrial, and ventricular cell types from the genomically edited
hESCs in order to study the effects of these mutations on the different cardiac lineages.
By studying the impact of various drug treatments, we can determine whether and how
the atrial remodeling and AF induced by GATA6 mutations relate to the remodeling
induced by AF itself. If the mechanisms are quite different, these studies could be used to
identify personalized optimal drug therapies for these patients carrying these mutations,
as well as numerous other patients who may have as yet unidentified mutations. Most
importantly, this type of investigation has the potential to generate and test new
hypotheses regarding propensity to and mechanism of disease that may be potential
preventative and/or therapeutic targets.
96
Other gene mutations will also be studied such as the TBX5 gene which causes Holt-
Oram syndrome and has been linked to lone AF 290,291. TBX5 likely lies upstream of
ANF and Cx-40, and may contribute to Nav1.5 expression 290,291. TBX5 mutations have
been associated with sinus node disease, AV node disease as well as QT prolongation 292.
Most recently, TBX5 has been identified as a novel locus for AF in a population of over
7000 European and Japanese patients with lone AF293. Although informative, the
analysis of isolated mutations in the backbone of an otherwise genetically normal stem
cell line will require complementary studies using hiPSCs. The strength in the hiPSC
model lies in the fact that in encompasses a host of unknown complementary and
confounding genetic variations which may play an indispensible role in the expression of
the disease experienced by the patients.
The final future direction to be discussed is the incorporation of bioengineering into our
AF model, that will use engineered heart tissue for analyses. Although the 2D geometry
of atrial cell sheets mimic in many respects that of atrial tissue, which is very thin and
unlikely to support 3D rotors (scroll waves), there are unwanted edge effects which
impact on source-sink matching, and have no parallels in human atria. Our atrial cell
sheets also show characteristics consistent with immaturity such as spontaneous beating.
Although it is conceivable that our cell sheets will mature with extended time in culture,
we have found limitations to time dependent maturation, consistent with other similar
studies153,244,294. In fact, studies of heart development show us that cardiac cells require
many environmental cues to drive adult patterning. Other efforts to promote maturity
have focused on electrical stimulation, mechanical strain, biochemical agents, co-
cultures, 3D culture, extracellular substrates, and genetic manipulations224. We plan to
97
study the impact of geometry and immaturity simultaneously using co-cultures of cardiac
fibroblasts (as previously described), as well as alternative sources of human cardiac
fibroblasts and CD90+ cells mixed with cardiomyocytes and incorporated into 3D
biowires295. Biowires can be mechanically loaded, electrically stimulated, and exposed to
selected chemical inducers of maturation. We will study biowires in detail using the same
approaches previously outlined including arrhythmia induction (i.e. complete electrical
datasets, beating rates, various membrane currents). We know that atrial cell sheets
support sustained rotors much more easily than ventricular cell sheets, and expect that
sustained rotors will be readily induced in atrial biowires. We will screen using markers
of structural maturation such as sarcomeric banding and myofilament density on
immunofluorescence and electron microscopy, physiologic maturation by examining the
amplitude of calcium transients and the shape and resting membrane potentials of
optically mapped and micropipette recorded action potentials, and finally looking at
effects on cell cycle activity as well as monitoring mRNA and protein levels 296 297 296.
Numerous studies have shown that cell responses to drugs in 3D culture better predict in
vivo tissue functionality compared to 2D high throughput screens. On the other hand,
high throughput drug screening technologies relying on single cells or looking at single
biochemical or gene expression assays have been enormously successful in testing and
developing novel compounds cheaply and quickly. Most recently there has been a
movement towards tissue based assays, or organ-on-chip platforms for high throughput
assays to improve the translational relevance of these screens by approximating more
closely the disease or the organ of interest. We would propose that our model is well
98
suited for the application of high throughput screening technologies once we can generate
tissue efficiently and reproducibly that has predictable performance parameters.
The strength in our model is the extreme nature of control which can ultimately be
exerted over countless variables, from the genetic make-up of the cell to the cell’s
environment. The challenge henceforth will be to understand each of these variables
under a defined set of extreme conditions such that the model’s behavior can be
understood and evaluated. A further challenge will be the scalability of the readout
analysis which must be interpreted in quantifiable patterns, and with the assistance of
computer learning, rather than the current form of qualitative observations and semi-
quantitative recordings that are hypothesis generating. The experiments described in our
future directions will define the most reliable measures, and help to focus our attention on
the outputs with the greatest signal. Once identified, and before a major investment into
scalability, these variables must be strictly correlated with important clinical endpoints to
ensure that our model has the capacity to exert a positive impact on patients for whom
these studies have been devised.
99
References
1 Andrade, J., Khairy, P., Dobrev, D. & Nattel, S. The clinical profile and
pathophysiology of atrial fibrillation: relationships among clinical features,
epidemiology, and mechanisms. Circulation research 114, 1453-1468,
doi:10.1161/CIRCRESAHA.114.303211 (2014).
2 Lloyd-Jones, D. M. et al. Lifetime risk for development of atrial fibrillation: the
Framingham Heart Study. Circulation 110, 1042-1046,
doi:10.1161/01.CIR.0000140263.20897.42 (2004).
3 Weiss, J. N. et al. The dynamics of cardiac fibrillation. Circulation 112, 1232-
1240, doi:10.1161/CIRCULATIONAHA.104.529545 (2005).
4 McManus, D. D., Rienstra, M. & Benjamin, E. J. An update on the prognosis of
patients with atrial fibrillation. Circulation 126, e143-146,
doi:10.1161/CIRCULATIONAHA.112.129759 (2012).
5 Heeringa, J. et al. Subclinical atherosclerosis and risk of atrial fibrillation: the
rotterdam study. Archives of internal medicine 167, 382-387, doi:167/4/382
[pii]
10.1001/archinte.167.4.382 (2007).
6 Mathew, S. T., Patel, J. & Joseph, S. Atrial fibrillation: mechanistic insights and
treatment options. European journal of internal medicine 20, 672-681,
doi:S0953-6205(09)00140-X [pii]
10.1016/j.ejim.2009.07.011 (2009).
7 Movahed, M. R., Hashemzadeh, M. & Jamal, M. M. Diabetes mellitus is a strong,
independent risk for atrial fibrillation and flutter in addition to other
cardiovascular disease. Int J Cardiol 105, 315-318, doi:S0167-
5273(05)00523-1 [pii]
10.1016/j.ijcard.2005.02.050 (2005).
8 Nattel, S. Therapeutic implications of atrial fibrillation mechanisms: can
mechanistic insights be used to improve AF management? Cardiovascular
research 54, 347-360 (2002).
9 Wodchis, W. P., Bhatia, R. S., Leblanc, K., Meshkat, N. & Morra, D. A review of
the cost of atrial fibrillation. Value in health : the journal of the International
Society for Pharmacoeconomics and Outcomes Research 15, 240-248,
doi:10.1016/j.jval.2011.09.009 (2012).
10 Khaykin, Y. & Shamiss, Y. Cost of atrial fibrillation: invasive vs non-invasive
management in 2012. Current cardiology reviews 8, 368-373 (2012).
11 Ganesan, A. N. et al. Long-term outcomes of catheter ablation of atrial
fibrillation: a systematic review and meta-analysis. Journal of the American
Heart Association 2, e004549, doi:10.1161/JAHA.112.004549 (2013).
12 Mandapati, R., Skanes, A., Chen, J., Berenfeld, O. & Jalife, J. Stable
microreentrant sources as a mechanism of atrial fibrillation in the isolated
sheep heart. Circulation 101, 194-199 (2000).
13 Sarmast, F. et al. Cholinergic atrial fibrillation: I(K,ACh) gradients determine
unequal left/right atrial frequencies and rotor dynamics. Cardiovascular
research 59, 863-873 (2003).
100
14 Nattel, S., Li, D. & Yue, L. Basic mechanisms of atrial fibrillation--very new
insights into very old ideas. Annu Rev Physiol 62, 51-77 (2000).
15 Schotten, U. et al. Electrical and contractile remodeling during the first days
of atrial fibrillation go hand in hand. Circulation 107, 1433-1439 (2003).
16 Shinagawa, K., Shi, Y. F., Tardif, J. C., Leung, T. K. & Nattel, S. Dynamic nature
of atrial fibrillation substrate during development and reversal of heart
failure in dogs. Circulation 105, 2672-2678 (2002).
17 Farrar, M. W., Bogart, D. B., Chapman, S. S. & Rigden, L. B. Atrial fibrillation in
athletes. Mo Med 103, 297-301 (2006).
18 Nattel, S. & Opie, L. H. Controversies in atrial fibrillation. Lancet 367, 262-272
(2006).
19 January, C. T. et al. 2014 AHA/ACC/HRS guideline for the management of
patients with atrial fibrillation: a report of the American College of
Cardiology/American Heart Association Task Force on Practice Guidelines
and the Heart Rhythm Society. Journal of the American College of Cardiology
64, e1-76, doi:10.1016/j.jacc.2014.03.022 (2014).
20 Hohnloser, S. H., Kuck, K. H. & Lilienthal, J. Rhythm or rate control in atrial
fibrillation--Pharmacological Intervention in Atrial Fibrillation (PIAF): a
randomised trial. Lancet 356, 1789-1794 (2000).
21 Wyse, D. G. et al. A comparison of rate control and rhythm control in patients
with atrial fibrillation. The New England journal of medicine 347, 1825-1833,
doi:10.1056/NEJMoa021328 (2002).
22 Van Gelder, I. C. et al. A comparison of rate control and rhythm control in
patients with recurrent persistent atrial fibrillation. The New England journal
of medicine 347, 1834-1840, doi:10.1056/NEJMoa021375 (2002).
23 Carlsson, J. et al. Randomized trial of rate-control versus rhythm-control in
persistent atrial fibrillation: the Strategies of Treatment of Atrial Fibrillation
(STAF) study. Journal of the American College of Cardiology 41, 1690-1696
(2003).
24 Caldeira, D., David, C. & Sampaio, C. Rate vs rhythm control in patients with
atrial fibrillation and heart failure: a systematic review and meta-analysis of
randomised controlled trials. European journal of internal medicine 22, 448-
455, doi:10.1016/j.ejim.2011.05.001 (2011).
25 Roy, D. et al. Rhythm control versus rate control for atrial fibrillation and
heart failure. The New England journal of medicine 358, 2667-2677,
doi:10.1056/NEJMoa0708789 (2008).
26 Corley, S. D. et al. Relationships between sinus rhythm, treatment, and
survival in the Atrial Fibrillation Follow-Up Investigation of Rhythm
Management (AFFIRM) Study. Circulation 109, 1509-1513,
doi:10.1161/01.CIR.0000121736.16643.11 (2004).
27 Singh, S. N. et al. Quality of life and exercise performance in patients in sinus
rhythm versus persistent atrial fibrillation: a Veterans Affairs Cooperative
Studies Program Substudy. Journal of the American College of Cardiology 48,
721-730, doi:10.1016/j.jacc.2006.03.051 (2006).
28 Guedon-Moreau, L. et al. Impact of the control of symptomatic paroxysmal
atrial fibrillation on health-related quality of life. Europace : European pacing,
101
arrhythmias, and cardiac electrophysiology : journal of the working groups on
cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the
European Society of Cardiology 12, 634-642, doi:10.1093/europace/euq007
(2010).
29 Opolski, G. et al. Rate control vs rhythm control in patients with nonvalvular
persistent atrial fibrillation: the results of the Polish How to Treat Chronic
Atrial Fibrillation (HOT CAFE) Study. Chest 126, 476-486,
doi:10.1378/chest.126.2.476 (2004).
30 Dorian, P. & Mangat, I. Quality of life variables in the selection of rate versus
rhythm control in patients with atrial fibrillation: observations from the
Canadian Trial of Atrial Fibrillation. Cardiac electrophysiology review 7, 276-
279, doi:10.1023/B:CEPR.0000012395.33292.cd (2003).
31 Nishida, K. & Nattel, S. Atrial fibrillation compendium: historical context and
detailed translational perspective on an important clinical problem.
Circulation research 114, 1447-1452, doi:10.1161/CIRCRESAHA.114.303466
(2014).
32 Haissaguerre, M. et al. Spontaneous initiation of atrial fibrillation by ectopic
beats originating in the pulmonary veins. The New England journal of
medicine 339, 659-666, doi:10.1056/NEJM199809033391003 (1998).
33 Scheinman, M. M. & Morady, F. Nonpharmacological approaches to atrial
fibrillation. Circulation 103, 2120-2125 (2001).
34 Nishida, K., Datino, T., Macle, L. & Nattel, S. Atrial fibrillation ablation:
translating basic mechanistic insights to the patient. Journal of the American
College of Cardiology 64, 823-831, doi:10.1016/j.jacc.2014.06.1172 (2014).
35 Chang, S. L. et al. Biatrial substrate properties in patients with atrial
fibrillation. Journal of cardiovascular electrophysiology 18, 1134-1139,
doi:10.1111/j.1540-8167.2007.00941.x (2007).
36 Pappone, C. et al. Mortality, morbidity, and quality of life after circumferential
pulmonary vein ablation for atrial fibrillation: outcomes from a controlled
nonrandomized long-term study. Journal of the American College of
Cardiology 42, 185-197 (2003).
37 Gaita, F. et al. Long-term clinical results of 2 different ablation strategies in
patients with paroxysmal and persistent atrial fibrillation. Circulation.
Arrhythmia and electrophysiology 1, 269-275,
doi:10.1161/CIRCEP.108.774885 (2008).
38 Bhargava, M. et al. Impact of type of atrial fibrillation and repeat catheter
ablation on long-term freedom from atrial fibrillation: results from a
multicenter study. Heart rhythm : the official journal of the Heart Rhythm
Society 6, 1403-1412, doi:10.1016/j.hrthm.2009.06.014 (2009).
39 Tzou, W. S. et al. Long-term outcome after successful catheter ablation of
atrial fibrillation. Circulation. Arrhythmia and electrophysiology 3, 237-242,
doi:10.1161/CIRCEP.109.923771 (2010).
40 Ouyang, F. et al. Long-term results of catheter ablation in paroxysmal atrial
fibrillation: lessons from a 5-year follow-up. Circulation 122, 2368-2377,
doi:10.1161/CIRCULATIONAHA.110.946806 (2010).
102
41 Medi, C. et al. Pulmonary vein antral isolation for paroxysmal atrial
fibrillation: results from long-term follow-up. Journal of cardiovascular
electrophysiology 22, 137-141, doi:10.1111/j.1540-8167.2010.01885.x
(2011).
42 Wijffels, M. C., Kirchhof, C. J., Dorland, R. & Allessie, M. A. Atrial fibrillation
begets atrial fibrillation. A study in awake chronically instrumented goats.
Circulation 92, 1954-1968 (1995).
43 Anne, W. et al. Self-terminating AF depends on electrical remodeling while
persistent AF depends on additional structural changes in a rapid atrially
paced sheep model. Journal of molecular and cellular cardiology 43, 148-158
(2007).
44 Nishida, K., Michael, G., Dobrev, D. & Nattel, S. Animal models for atrial
fibrillation: clinical insights and scientific opportunities. Europace : European
pacing, arrhythmias, and cardiac electrophysiology : journal of the working
groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology
of the European Society of Cardiology 12, 160-172, doi:eup328 [pii]
10.1093/europace/eup328 (2010).
45 Amin, A. S., Tan, H. L. & Wilde, A. A. Cardiac ion channels in health and
disease. Heart rhythm : the official journal of the Heart Rhythm Society 7, 117-
126, doi:10.1016/j.hrthm.2009.08.005 (2010).
46 Tsai, C. T., Lai, L. P., Hwang, J. J., Lin, J. L. & Chiang, F. T. Molecular genetics of
atrial fibrillation. Journal of the American College of Cardiology 52, 241-250,
doi:S0735-1097(08)01200-X [pii]
10.1016/j.jacc.2008.02.072 (2008).
47 Choi, G. et al. Spectrum and frequency of cardiac channel defects in
swimming-triggered arrhythmia syndromes. Circulation 110, 2119-2124,
doi:01.CIR.0000144471.98080.CA [pii]
10.1161/01.CIR.0000144471.98080.CA (2004).
48 Watanabe, H. et al. Mutations in sodium channel beta1- and beta2-subunits
associated with atrial fibrillation. Circulation. Arrhythmia and
electrophysiology 2, 268-275, doi:CIRCEP.108.779181 [pii]
10.1161/CIRCEP.108.779181 (2009).
49 Francis, J. & Antzelevitch, C. Atrial fibrillation and Brugada syndrome. Journal
of the American College of Cardiology 51, 1149-1153, doi:S0735-
1097(08)00174-5 [pii]
10.1016/j.jacc.2007.10.062 (2008).
50 Chen, Y. H. et al. KCNQ1 gain-of-function mutation in familial atrial
fibrillation. Science 299, 251-254 (2003).
51 Dobrev, D. et al. The G protein-gated potassium current I(K,ACh) is
constitutively active in patients with chronic atrial fibrillation. Circulation
112, 3697-3706 (2005).
52 Li, J., McLerie, M. & Lopatin, A. N. Transgenic upregulation of IK1 in the
mouse heart leads to multiple abnormalities of cardiac excitability. American
journal of physiology. Heart and circulatory physiology 287, H2790-2802
(2004).
103
53 Roberts, J. D. & Gollob, M. H. Impact of genetic discoveries on the
classification of lone atrial fibrillation. Journal of the American College of
Cardiology 55, 705-712, doi:S0735-1097(09)03989-8 [pii]
10.1016/j.jacc.2009.12.005 [doi].
54 Temple, J. et al. Atrial fibrillation in KCNE1-null mice. Circulation research 97,
62-69 (2005).
55 Fareh, S., Villemaire, C. & Nattel, S. Importance of refractoriness
heterogeneity in the enhanced vulnerability to atrial fibrillation induction
caused by tachycardia-induced atrial electrical remodeling. Circulation 98,
2202-2209 (1998).
56 Lukas A & Antzelevitch C. Phase 2 reentry as a mechanism of initiation of
circus movement reentry in canine epicardium exposed to simulated
ischemia. Cardiovascular research 32, 593-603 (1996).
57 Katsouras, G. et al. Differences in atrial fibrillation properties under vagal
nerve stimulation versus atrial tachycardia remodeling. Heart rhythm : the
official journal of the Heart Rhythm Society 6, 1465-1472 (2009).
58 Wang, Z., Page, P. & Nattel, S. Mechanism of flecainide's antiarrhythmic action
in experimental atrial fibrillation. Circulation research 71, 271-287 (1992).
59 Kneller, J. et al. Cholinergic atrial fibrillation in a computer model of a two-
dimensional sheet of canine atrial cells with realistic ionic properties.
Circulation research 90, E73-87 (2002).
60 Korantzopoulos, P., Kolettis, T. M., Goudevenos, J. A. & Siogas, K. Errors and
pitfalls in the non-invasive management of atrial fibrillation. Int J Cardiol
104, 125-130, doi:S0167-5273(05)00210-X [pii]
10.1016/j.ijcard.2004.11.014 (2005).
61 Po, S. S. et al. Rapid and stable re-entry within the pulmonary vein as a
mechanism initiating paroxysmal atrial fibrillation. Journal of the American
College of Cardiology 45, 1871-1877 (2005).
62 Desplantez, T., Dupont, E., Severs, N. J. & Weingart, R. Gap junction channels
and cardiac impulse propagation. The Journal of membrane biology 218, 13-
28, doi:10.1007/s00232-007-9046-8 (2007).
63 Henriquez, A. P. et al. Influence of dynamic gap junction resistance on
impulse propagation in ventricular myocardium: a computer simulation
study. Biophysical journal 81, 2112-2121, doi:S0006-3495(01)75859-6 [pii]
10.1016/S0006-3495(01)75859-6 (2001).
64 Hinch, R. An analytical study of the physiology and pathology of the
propagation of cardiac action potentials. Prog Biophys Mol Biol 78, 45-81,
doi:S0079610702000068 [pii] (2002).
65 Amin, A. S., Asghari-Roodsari, A. & Tan, H. L. Cardiac sodium channelopathies.
Pflugers Archiv : European journal of physiology 460, 223-237,
doi:10.1007/s00424-009-0761-0 (2010).
66 van der Velden, H. M. & Jongsma, H. J. Cardiac gap junctions and connexins:
their role in atrial fibrillation and potential as therapeutic targets.
Cardiovascular research 54, 270-279 (2002).
104
67 van der Velden, H. M. et al. Gap junctional remodeling in relation to
stabilization of atrial fibrillation in the goat. Cardiovascular research 46, 476-
486 (2000).
68 Verheule, S. et al. Cardiac conduction abnormalities in mice lacking the gap
junction protein connexin40. Journal of cardiovascular electrophysiology 10,
1380-1389 (1999).
69 Gollob, M. H. et al. Somatic mutations in the connexin 40 gene (GJA5) in atrial
fibrillation. The New England journal of medicine 354, 2677-2688,
doi:10.1056/NEJMoa052800 (2006).
70 Burstein, B. et al. Changes in connexin expression and the atrial fibrillation
substrate in congestive heart failure. Circulation research 105, 1213-1222,
doi:CIRCRESAHA.108.183400 [pii]
10.1161/CIRCRESAHA.108.183400 (2009).
71 Chaldoupi, S. M., Loh, P., Hauer, R. N., de Bakker, J. M. & van Rijen, H. V. The
role of connexin40 in atrial fibrillation. Cardiovascular research 84, 15-23,
doi:cvp203 [pii]
10.1093/cvr/cvp203 (2009).
72 Liu, X. et al. Decreased connexin 43 and increased fibrosis in atrial regions
susceptible to complex fractionated atrial electrograms. Cardiology 114, 22-
29, doi:000210398 [pii]
10.1159/000210398 (2009).
73 Rossman, E. I. et al. The gap junction modifier, GAP-134 [(2S,4R)-1-(2-
aminoacetyl)-4-benzamido-pyrrolidine-2-carboxylic acid], improves
conduction and reduces atrial fibrillation/flutter in the canine sterile
pericarditis model. The Journal of pharmacology and experimental
therapeutics 329, 1127-1133, doi:jpet.108.150102 [pii]
10.1124/jpet.108.150102 (2009).
74 Li, J. Y. et al. Atrial gap junctions, NF-kappaB and fibrosis in patients
undergoing coronary artery bypass surgery: the relationship with
postoperative atrial fibrillation. Cardiology 112, 81-88, doi:000141012 [pii]
10.1159/000141012 (2009).
75 Rasmussen, H. S., Allen, M. J., Blackburn, K. J., Butrous, G. S. & Dalrymple, H. W.
Dofetilide, a novel class III antiarrhythmic agent. Journal of cardiovascular
pharmacology 20 Suppl 2, S96-105 (1992).
76 Gwilt, M. et al. UK-68,798: a novel, potent and highly selective class III
antiarrhythmic agent which blocks potassium channels in cardiac cells. The
Journal of pharmacology and experimental therapeutics 256, 318-324 (1991).
77 Knilans, T. K., Lathrop, D. A., Nanasi, P. P., Schwartz, A. & Varro, A. Rate and
concentration-dependent effects of UK-68,798, a potent new class III
antiarrhythmic, on canine Purkinje fibre action potential duration and Vmax.
British journal of pharmacology 103, 1568-1572 (1991).
78 Singh, S. et al. Efficacy and safety of oral dofetilide in converting to and
maintaining sinus rhythm in patients with chronic atrial fibrillation or atrial
flutter: the symptomatic atrial fibrillation investigative research on dofetilide
(SAFIRE-D) study. Circulation 102, 2385-2390 (2000).
105
79 Kober, L. et al. Effect of dofetilide in patients with recent myocardial
infarction and left-ventricular dysfunction: a randomised trial. Lancet 356,
2052-2058 (2000).
80 Torp-Pedersen, C. et al. Dofetilide in patients with congestive heart failure
and left ventricular dysfunction. Danish Investigations of Arrhythmia and
Mortality on Dofetilide Study Group. The New England journal of medicine
341, 857-865, doi:10.1056/NEJM199909163411201 (1999).
81 Fuster, V. et al. 2011 ACCF/AHA/HRS focused updates incorporated into the
ACC/AHA/ESC 2006 Guidelines for the management of patients with atrial
fibrillation: a report of the American College of Cardiology
Foundation/American Heart Association Task Force on Practice Guidelines
developed in partnership with the European Society of Cardiology and in
collaboration with the European Heart Rhythm Association and the Heart
Rhythm Society. Journal of the American College of Cardiology 57, e101-198,
doi:10.1016/j.jacc.2010.09.013 (2011).
82 Sanoski, C. A. Atrial fibrillation and managed care: current approaches and
future directions for long-term therapy. Introduction. Journal of managed
care pharmacy : JMCP 15, S3 (2009).
83 Campbell, T. J. Kinetics of onset of rate-dependent effects of Class I
antiarrhythmic drugs are important in determining their effects on
refractoriness in guinea-pig ventricle, and provide a theoretical basis for
their subclassification. Cardiovascular research 17, 344-352 (1983).
84 Harrison, D. C. Antiarrhythmic drug classification: new science and practical
applications. The American journal of cardiology 56, 185-187 (1985).
85 Wang, Z., Feng, J. & Nattel, S. Idiopathic atrial fibrillation in dogs:
electrophysiologic determinants and mechanisms of antiarrhythmic action of
flecainide. Journal of the American College of Cardiology 26, 277-286 (1995).
86 Anderson, J. L. et al. Prevention of symptomatic recurrences of paroxysmal
atrial fibrillation in patients initially tolerating antiarrhythmic therapy. A
multicenter, double-blind, crossover study of flecainide and placebo with
transtelephonic monitoring. Flecainide Supraventricular Tachycardia Study
Group. Circulation 80, 1557-1570 (1989).
87 Pietersen, A. H. & Hellemann, H. Usefulness of flecainide for prevention of
paroxysmal atrial fibrillation and flutter. Danish-Norwegian Flecainide
Multicenter Study Group. The American journal of cardiology 67, 713-717
(1991).
88 Akiyama, T., Pawitan, Y., Greenberg, H., Kuo, C. S. & Reynolds-Haertle, R. A.
Increased risk of death and cardiac arrest from encainide and flecainide in
patients after non-Q-wave acute myocardial infarction in the Cardiac
Arrhythmia Suppression Trial. CAST Investigators. The American journal of
cardiology 68, 1551-1555 (1991).
89 Nappi, J. M. & Anderson, J. L. Flecainide: a new prototype antiarrhythmic
agent. Pharmacotherapy 5, 209-221 (1985).
90 Estes, N. A., 3rd, Garan, H. & Ruskin, J. N. Electrophysiologic properties of
flecainide acetate. The American journal of cardiology 53, 26B-29B (1984).
106
91 Packer, D. L., Munger, T. M., Johnson, S. B. & Cragun, K. T. Mechanism of lethal
proarrhythmia observed in the Cardiac Arrhythmia Suppression Trial: role of
adrenergic modulation of drug binding. Pacing and clinical electrophysiology :
PACE 20, 455-467 (1997).
92 Pandit, S. V. & Jalife, J. Rotors and the dynamics of cardiac fibrillation.
Circulation research 112, 849-862, doi:10.1161/CIRCRESAHA.111.300158
(2013).
93 Mines, G. R. On dynamic equilibrium in the heart. The Journal of physiology
46, 349-383 (1913).
94 Lewis, T. Oliver-Sharpey Lectures ON THE NATURE OF FLUTTER AND
FIBRILLATION OF THE AURICLE. British medical journal 1, 590-593 (1921).
95 Allessie, M. A., Bonke, F. I. & Schopman, F. J. Circus movement in rabbit atrial
muscle as a mechanism of trachycardia. Circulation research 33, 54-62
(1973).
96 Garanin, O. A., Popov, N. V., Efimov, S. V., Udovikov, A. I. & Grigor'eva, G. V.
[The dynamics of the species composition of nidiculous animals in different
types of nests of the little suslik]. Parazitologiia 29, 37-42 (1995).
97 Zhabotinsky, A. M. A history of chemical oscillations and waves. Chaos 1, 379-
386, doi:10.1063/1.165848 (1991).
98 Vaquero, M., Calvo, D. & Jalife, J. Cardiac fibrillation: from ion channels to
rotors in the human heart. Heart rhythm : the official journal of the Heart
Rhythm Society 5, 872-879, doi:10.1016/j.hrthm.2008.02.034 (2008).
99 Gray, R. A., Pertsov, A. M. & Jalife, J. Spatial and temporal organization during
cardiac fibrillation. Nature 392, 75-78, doi:10.1038/32164 (1998).
100 Fast, V. G. & Kleber, A. G. Role of wavefront curvature in propagation of
cardiac impulse. Cardiovascular research 33, 258-271 (1997).
101 Panfilov, A. V. & Pertsov, A. M. [Mechanism of spiral waves initiation in active
media, connected with critical curvature phenomenon]. Biofizika 27, 886-889
(1982).
102 Schotten, U., Verheule, S., Kirchhof, P. & Goette, A. Pathophysiological
mechanisms of atrial fibrillation: a translational appraisal. Physiological
reviews 91, 265-325, doi:10.1152/physrev.00031.2009 (2011).
103 Gray, R. A. et al. Mechanisms of cardiac fibrillation. Science 270, 1222-1223;
author reply 1224-1225 (1995).
104 Jalife, J. Inward rectifier potassium channels control rotor frequency in
ventricular fibrillation. Heart rhythm : the official journal of the Heart Rhythm
Society 6, S44-48, doi:10.1016/j.hrthm.2009.07.019 (2009).
105 Jalife, J. & Pandit, S. V. Ionic mechanisms of wavebreak in fibrillation. Heart
rhythm : the official journal of the Heart Rhythm Society 2, 660-663,
doi:10.1016/j.hrthm.2005.03.016 (2005).
106 Pertsov, A. M., Davidenko, J. M., Salomonsz, R., Baxter, W. T. & Jalife, J. Spiral
waves of excitation underlie reentrant activity in isolated cardiac muscle.
Circulation research 72, 631-650 (1993).
107 Wohlfart, B. Analysis of mechanical alternans in rabbit papillary muscle. Acta
physiologica Scandinavica 115, 405-414, doi:10.1111/j.1748-
1716.1982.tb07098.x (1982).
107
108 Walker, M. L. & Rosenbaum, D. S. Repolarization alternans: implications for
the mechanism and prevention of sudden cardiac death. Cardiovascular
research 57, 599-614 (2003).
109 Pitt, G. S., Dun, W. & Boyden, P. A. Remodeled cardiac calcium channels.
Journal of molecular and cellular cardiology 41, 373-388,
doi:10.1016/j.yjmcc.2006.06.071 (2006).
110 Verrier, R. L. & Nieminen, T. T-wave alternans as a therapeutic marker for
antiarrhythmic agents. Journal of cardiovascular pharmacology 55, 544-554,
doi:10.1097/FJC.0b013e3181d6b781 (2010).
111 Nolasco, J. B. & Dahlen, R. W. A graphic method for the study of alternation in
cardiac action potentials. Journal of applied physiology 25, 191-196 (1968).
112 Tolkacheva, E. G., Anumonwo, J. M. & Jalife, J. Action potential duration
restitution portraits of mammalian ventricular myocytes: role of calcium
current. Biophysical journal 91, 2735-2745,
doi:10.1529/biophysj.106.083865 (2006).
113 Tolkacheva, E. G., Romeo, M. M., Guerraty, M. & Gauthier, D. J. Condition for
alternans and its control in a two-dimensional mapping model of paced
cardiac dynamics. Physical review. E, Statistical, nonlinear, and soft matter
physics 69, 031904 (2004).
114 Watanabe, M. A. & Koller, M. L. Mathematical analysis of dynamics of cardiac
memory and accommodation: theory and experiment. American journal of
physiology. Heart and circulatory physiology 282, H1534-1547,
doi:10.1152/ajpheart.00351.2001 (2002).
115 Gelzer, A. R. et al. Dynamic mechanism for initiation of ventricular fibrillation
in vivo. Circulation 118, 1123-1129,
doi:10.1161/CIRCULATIONAHA.107.738013 (2008).
116 Kanaporis, G. & Blatter, L. A. The mechanisms of calcium cycling and action
potential dynamics in cardiac alternans. Circulation research 116, 846-856,
doi:10.1161/CIRCRESAHA.116.305404 (2015).
117 Herweg, B., Kowalski, M. & Steinberg, J. S. Termination of persistent atrial
fibrillation resistant to cardioversion by a single radiofrequency application.
Pacing and clinical electrophysiology : PACE 26, 1420-1423 (2003).
118 Lazar, S., Dixit, S., Marchlinski, F. E., Callans, D. J. & Gerstenfeld, E. P. Presence
of left-to-right atrial frequency gradient in paroxysmal but not persistent
atrial fibrillation in humans. Circulation 110, 3181-3186,
doi:10.1161/01.CIR.0000147279.91094.5E (2004).
119 Wu, T. J. et al. Simultaneous biatrial computerized mapping during
permanent atrial fibrillation in patients with organic heart disease. Journal of
cardiovascular electrophysiology 13, 571-577 (2002).
120 Sahadevan, J. et al. Epicardial mapping of chronic atrial fibrillation in
patients: preliminary observations. Circulation 110, 3293-3299,
doi:10.1161/01.CIR.0000147781.02738.13 (2004).
121 Sanders, P. et al. Spectral analysis identifies sites of high-frequency activity
maintaining atrial fibrillation in humans. Circulation 112, 789-797,
doi:10.1161/CIRCULATIONAHA.104.517011 (2005).
108
122 Schricker, A. A., Lalani, G. G., Krummen, D. E. & Narayan, S. M. Rotors as
drivers of atrial fibrillation and targets for ablation. Current cardiology
reports 16, 509, doi:10.1007/s11886-014-0509-0 (2014).
123 Ganesan, A. N. et al. Bipolar electrogram shannon entropy at sites of
rotational activation: implications for ablation of atrial fibrillation.
Circulation. Arrhythmia and electrophysiology 6, 48-57,
doi:10.1161/CIRCEP.112.976654 (2013).
124 Narayan, S. M. et al. Ablation of rotor and focal sources reduces late
recurrence of atrial fibrillation compared with trigger ablation alone:
extended follow-up of the CONFIRM trial (Conventional Ablation for Atrial
Fibrillation With or Without Focal Impulse and Rotor Modulation). Journal of
the American College of Cardiology 63, 1761-1768,
doi:10.1016/j.jacc.2014.02.543 (2014).
125 Miller, J. M. et al. Initial independent outcomes from focal impulse and rotor
modulation ablation for atrial fibrillation: multicenter FIRM registry. Journal
of cardiovascular electrophysiology 25, 921-929, doi:10.1111/jce.12474
(2014).
126 Thomson, J. A. et al. Embryonic stem cell lines derived from human
blastocysts. Science 282, 1145-1147 (1998).
127 Takahashi, K. & Yamanaka, S. Induction of pluripotent stem cells from mouse
embryonic and adult fibroblast cultures by defined factors. Cell 126, 663-
676, doi:10.1016/j.cell.2006.07.024 (2006).
128 Takahashi, K., Okita, K., Nakagawa, M. & Yamanaka, S. Induction of
pluripotent stem cells from fibroblast cultures. Nature protocols 2, 3081-
3089, doi:10.1038/nprot.2007.418 (2007).
129 Murry, C. E. & Keller, G. Differentiation of embryonic stem cells to clinically
relevant populations: lessons from embryonic development. Cell 132, 661-
680, doi:10.1016/j.cell.2008.02.008 (2008).
130 Ng, E. S. et al. The primitive streak gene Mixl1 is required for efficient
haematopoiesis and BMP4-induced ventral mesoderm patterning in
differentiating ES cells. Development 132, 873-884, doi:10.1242/dev.01657
(2005).
131 Kubo, A. et al. Development of definitive endoderm from embryonic stem
cells in culture. Development 131, 1651-1662, doi:10.1242/dev.01044
(2004).
132 Wiles, M. V. & Johansson, B. M. Embryonic stem cell development in a
chemically defined medium. Experimental cell research 247, 241-248,
doi:10.1006/excr.1998.4353 (1999).
133 Yasunaga, M. et al. Induction and monitoring of definitive and visceral
endoderm differentiation of mouse ES cells. Nature biotechnology 23, 1542-
1550, doi:10.1038/nbt1167 (2005).
134 Fehling, H. J. et al. Tracking mesoderm induction and its specification to the
hemangioblast during embryonic stem cell differentiation. Development 130,
4217-4227 (2003).
135 Gadue, P., Huber, T. L., Paddison, P. J. & Keller, G. M. Wnt and TGF-beta
signaling are required for the induction of an in vitro model of primitive
109
streak formation using embryonic stem cells. Proceedings of the National
Academy of Sciences of the United States of America 103, 16806-16811,
doi:10.1073/pnas.0603916103 (2006).
136 Tada, S. et al. Characterization of mesendoderm: a diverging point of the
definitive endoderm and mesoderm in embryonic stem cell differentiation
culture. Development 132, 4363-4374, doi:10.1242/dev.02005 (2005).
137 Ying, Q. L., Stavridis, M., Griffiths, D., Li, M. & Smith, A. Conversion of
embryonic stem cells into neuroectodermal precursors in adherent
monoculture. Nature biotechnology 21, 183-186, doi:10.1038/nbt780 (2003).
138 Hogan, B. L. Bone morphogenetic proteins in development. Current opinion in
genetics & development 6, 432-438 (1996).
139 Conlon, F. L. et al. A primary requirement for nodal in the formation and
maintenance of the primitive streak in the mouse. Development 120, 1919-
1928 (1994).
140 Schier, A. F. Nodal signaling in vertebrate development. Annual review of cell
and developmental biology 19, 589-621,
doi:10.1146/annurev.cellbio.19.041603.094522 (2003).
141 Gadue, P., Huber, T. L., Nostro, M. C., Kattman, S. & Keller, G. M. Germ layer
induction from embryonic stem cells. Experimental hematology 33, 955-964,
doi:10.1016/j.exphem.2005.06.009 (2005).
142 Ema, M., Takahashi, S. & Rossant, J. Deletion of the selection cassette, but not
cis-acting elements, in targeted Flk1-lacZ allele reveals Flk1 expression in
multipotent mesodermal progenitors. Blood 107, 111-117,
doi:10.1182/blood-2005-05-1970 (2006).
143 Kataoka, H. et al. Expressions of PDGF receptor alpha, c-Kit and Flk1 genes
clustering in mouse chromosome 5 define distinct subsets of nascent
mesodermal cells. Development, growth & differentiation 39, 729-740 (1997).
144 Ueno, S. et al. Biphasic role for Wnt/beta-catenin signaling in cardiac
specification in zebrafish and embryonic stem cells. Proceedings of the
National Academy of Sciences of the United States of America 104, 9685-9690,
doi:10.1073/pnas.0702859104 (2007).
145 Era, T. et al. Multiple mesoderm subsets give rise to endothelial cells,
whereas hematopoietic cells are differentiated only from a restricted subset
in embryonic stem cell differentiation culture. Stem cells 26, 401-411,
doi:10.1634/stemcells.2006-0809 (2008).
146 Moretti, A. et al. Multipotent embryonic isl1+ progenitor cells lead to cardiac,
smooth muscle, and endothelial cell diversification. Cell 127, 1151-1165,
doi:10.1016/j.cell.2006.10.029 (2006).
147 Wu, S. M. et al. Developmental origin of a bipotential myocardial and smooth
muscle cell precursor in the mammalian heart. Cell 127, 1137-1150,
doi:10.1016/j.cell.2006.10.028 (2006).
148 DeRuiter, M. C., Poelmann, R. E., VanderPlas-de Vries, I., Mentink, M. M. &
Gittenberger-de Groot, A. C. The development of the myocardium and
endocardium in mouse embryos. Fusion of two heart tubes? Anatomy and
embryology 185, 461-473 (1992).
110
149 Lian, X. et al. Directed cardiomyocyte differentiation from human pluripotent
stem cells by modulating Wnt/beta-catenin signaling under fully defined
conditions. Nature protocols 8, 162-175, doi:10.1038/nprot.2012.150 (2013).
150 van den Heuvel, N. H., van Veen, T. A., Lim, B. & Jonsson, M. K. Lessons from
the heart: mirroring electrophysiological characteristics during cardiac
development to in vitro differentiation of stem cell derived cardiomyocytes.
Journal of molecular and cellular cardiology 67, 12-25,
doi:10.1016/j.yjmcc.2013.12.011 (2014).
151 Satin, J. et al. Calcium handling in human embryonic stem cell-derived
cardiomyocytes. Stem cells 26, 1961-1972, doi:10.1634/stemcells.2007-0591
(2008).
152 Itzhaki, I. et al. Calcium handling in human induced pluripotent stem cell
derived cardiomyocytes. PloS one 6, e18037,
doi:10.1371/journal.pone.0018037 (2011).
153 Lundy, S. D., Zhu, W. Z., Regnier, M. & Laflamme, M. A. Structural and
functional maturation of cardiomyocytes derived from human pluripotent
stem cells. Stem cells and development 22, 1991-2002,
doi:10.1089/scd.2012.0490 (2013).
154 Yang, L. et al. Human cardiovascular progenitor cells develop from a KDR+
embryonic-stem-cell-derived population. Nature 453, 524-528,
doi:10.1038/nature06894 (2008).
155 Kattman, S. J. et al. Stage-specific optimization of activin/nodal and BMP
signaling promotes cardiac differentiation of mouse and human pluripotent
stem cell lines. Cell stem cell 8, 228-240, doi:10.1016/j.stem.2010.12.008
(2011).
156 Ma, J. et al. High purity human-induced pluripotent stem cell-derived
cardiomyocytes: electrophysiological properties of action potentials and
ionic currents. American journal of physiology. Heart and circulatory
physiology 301, H2006-2017, doi:10.1152/ajpheart.00694.2011 (2011).
157 Xavier-Neto, J. et al. Retinoid signaling and cardiac anteroposterior
segmentation. Genesis 31, 97-104 (2001).
158 Hochgreb, T. et al. A caudorostral wave of RALDH2 conveys anteroposterior
information to the cardiac field. Development 130, 5363-5374,
doi:10.1242/dev.00750 (2003).
159 Xavier-Neto, J. et al. A retinoic acid-inducible transgenic marker of sino-atrial
development in the mouse heart. Development 126, 2677-2687 (1999).
160 Zhang, Q. et al. Direct differentiation of atrial and ventricular myocytes from
human embryonic stem cells by alternating retinoid signals. Cell research 21,
579-587, doi:10.1038/cr.2010.163 (2011).
161 Niederreither, K. et al. Embryonic retinoic acid synthesis is essential for heart
morphogenesis in the mouse. Development 128, 1019-1031 (2001).
162 Emoto, Y., Wada, H., Okamoto, H., Kudo, A. & Imai, Y. Retinoic acid-
metabolizing enzyme Cyp26a1 is essential for determining territories of
hindbrain and spinal cord in zebrafish. Developmental biology 278, 415-427,
doi:10.1016/j.ydbio.2004.11.023 (2005).
111
163 Lu, Z. Q., Sinha, A., Sharma, P., Kislinger, T. & Gramolini, A. O. Proteomic
analysis of human fetal atria and ventricle. Journal of proteome research 13,
5869-5878, doi:10.1021/pr5007685 (2014).
164 Ellinor, P. T. et al. Common variants in KCNN3 are associated with lone atrial
fibrillation. Nature genetics 42, 240-244, doi:10.1038/ng.537 (2010).
165 O'Brien, T. X., Lee, K. J. & Chien, K. R. Positional specification of ventricular
myosin light chain 2 expression in the primitive murine heart tube.
Proceedings of the National Academy of Sciences of the United States of
America 90, 5157-5161 (1993).
166 Bruneau, B. G. et al. Cardiac expression of the ventricle-specific homeobox
gene Irx4 is modulated by Nkx2-5 and dHand. Developmental biology 217,
266-277, doi:10.1006/dbio.1999.9548 (2000).
167 Moretti, A. et al. Patient-specific induced pluripotent stem-cell models for
long-QT syndrome. The New England journal of medicine 363, 1397-1409,
doi:10.1056/NEJMoa0908679 (2010).
168 Lan, F. et al. Abnormal calcium handling properties underlie familial
hypertrophic cardiomyopathy pathology in patient-specific induced
pluripotent stem cells. Cell stem cell 12, 101-113,
doi:10.1016/j.stem.2012.10.010 (2013).
169 Sun, N. et al. Patient-specific induced pluripotent stem cells as a model for
familial dilated cardiomyopathy. Science translational medicine 4, 130ra147,
doi:10.1126/scitranslmed.3003552 (2012).
170 Kim, C. et al. Studying arrhythmogenic right ventricular dysplasia with
patient-specific iPSCs. Nature 494, 105-110, doi:10.1038/nature11799
(2013).
171 Schwartz, P. J., Crotti, L. & Insolia, R. Long-QT syndrome: from genetics to
management. Circulation. Arrhythmia and electrophysiology 5, 868-877,
doi:10.1161/CIRCEP.111.962019 (2012).
172 Ackerman, M. J. et al. HRS/EHRA expert consensus statement on the state of
genetic testing for the channelopathies and cardiomyopathies: this document
was developed as a partnership between the Heart Rhythm Society (HRS)
and the European Heart Rhythm Association (EHRA). Europace : European
pacing, arrhythmias, and cardiac electrophysiology : journal of the working
groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology
of the European Society of Cardiology 13, 1077-1109,
doi:10.1093/europace/eur245 (2011).
173 Villain, E. et al. Low incidence of cardiac events with beta-blocking therapy in
children with long QT syndrome. European heart journal 25, 1405-1411,
doi:10.1016/j.ehj.2004.06.016 (2004).
174 Monnig, G. et al. Implantable cardioverter-defibrillator therapy in patients
with congenital long-QT syndrome: a long-term follow-up. Heart rhythm : the
official journal of the Heart Rhythm Society 2, 497-504,
doi:10.1016/j.hrthm.2005.02.008 (2005).
175 Schwartz, P. J. et al. Who are the long-QT syndrome patients who receive an
implantable cardioverter-defibrillator and what happens to them?: data from
the European Long-QT Syndrome Implantable Cardioverter-Defibrillator
112
(LQTS ICD) Registry. Circulation 122, 1272-1282,
doi:10.1161/CIRCULATIONAHA.110.950147 (2010).
176 Horner, J. M. et al. Implantable cardioverter defibrillator therapy for
congenital long QT syndrome: a single-center experience. Heart rhythm : the
official journal of the Heart Rhythm Society 7, 1616-1622,
doi:10.1016/j.hrthm.2010.08.023 (2010).
177 Schwartz, P. J. et al. Left cardiac sympathetic denervation in the management
of high-risk patients affected by the long-QT syndrome. Circulation 109,
1826-1833, doi:10.1161/01.CIR.0000125523.14403.1E
01.CIR.0000125523.14403.1E [pii] (2004).
178 Ackerman, M. J. et al. HRS/EHRA expert consensus statement on the state of
genetic testing for the channelopathies and cardiomyopathies: this document
was developed as a partnership between the Heart Rhythm Society (HRS)
and the European Heart Rhythm Association (EHRA). Europace 13, 1077-
1109, doi:eur245 [pii]
10.1093/europace/eur245 (2011).
179 Nunn, L. M. & Lambiase, P. D. Genetics and cardiovascular disease--causes
and prevention of unexpected sudden adult death: the role of the SADS clinic.
Heart 97, 1122-1127, doi:10.1136/hrt.2010.218511
hrt.2010.218511 [pii] (2011).
180 Zipes, D. P. et al. ACC/AHA/ESC 2006 guidelines for management of patients
with ventricular arrhythmias and the prevention of sudden cardiac death: a
report of the American College of Cardiology/American Heart Association
Task Force and the European Society of Cardiology Committee for Practice
Guidelines (Writing Committee to Develop Guidelines for Management of
Patients With Ventricular Arrhythmias and the Prevention of Sudden Cardiac
Death). J Am Coll Cardiol 48, e247-346 (2006).
181 Tang, A. S. et al. Canadian Cardiovascular Society/Canadian Heart Rhythm
Society position paper on implantable cardioverter defibrillator use in
Canada. Can J Cardiol 21 Suppl A, 11-18 (2005).
182 Olde Nordkamp, L. R. et al. The ICD for primary prevention in patients with
inherited cardiac diseases: indications, use, and outcome: a comparison with
secondary prevention. Circulation. Arrhythmia and electrophysiology 6, 91-
100, doi:10.1161/CIRCEP.112.975268 (2013).
183 Viswanathan, P. C. & Rudy, Y. Pause induced early afterdepolarizations in the
long QT syndrome: a simulation study. Cardiovascular research 42, 530-542
(1999).
184 Clancy, C. E. & Rudy, Y. Linking a genetic defect to its cellular phenotype in a
cardiac arrhythmia. Nature 400, 566-569, doi:10.1038/23034 (1999).
185 Lankipalli, R. S., Zhu, T., Guo, D. & Yan, G. X. Mechanisms underlying
arrhythmogenesis in long QT syndrome. Journal of electrocardiology 38, 69-
73, doi:10.1016/j.jelectrocard.2005.06.008 (2005).
186 Itzhaki, I. et al. Modelling the long QT syndrome with induced pluripotent
stem cells. Nature 471, 225-229, doi:10.1038/nature09747 (2011).
113
187 Yan, G. X. & Martin, J. Electrocardiographic T wave: a symbol of transmural
dispersion of repolarization in the ventricles. Journal of cardiovascular
electrophysiology 14, 639-640 (2003).
188 Yan, G. X. et al. Phase 2 early afterdepolarization as a trigger of polymorphic
ventricular tachycardia in acquired long-QT syndrome : direct evidence from
intracellular recordings in the intact left ventricular wall. Circulation 103,
2851-2856 (2001).
189 Yan, G. X. et al. Ventricular hypertrophy amplifies transmural repolarization
dispersion and induces early afterdepolarization. American journal of
physiology. Heart and circulatory physiology 281, H1968-1975 (2001).
190 Echt, D. S. et al. Mortality and morbidity in patients receiving encainide,
flecainide, or placebo. The Cardiac Arrhythmia Suppression Trial. The New
England journal of medicine 324, 781-788,
doi:10.1056/NEJM199103213241201 (1991).
191 Paul, S. M. et al. How to improve R&D productivity: the pharmaceutical
industry's grand challenge. Nature reviews. Drug discovery 9, 203-214,
doi:10.1038/nrd3078 (2010).
192 Kola, I. & Landis, J. Can the pharmaceutical industry reduce attrition rates?
Nature reviews. Drug discovery 3, 711-715, doi:10.1038/nrd1470 (2004).
193 MacDonald, J. S. & Robertson, R. T. Toxicity testing in the 21st century: a view
from the pharmaceutical industry. Toxicological sciences : an official journal
of the Society of Toxicology 110, 40-46, doi:10.1093/toxsci/kfp088 (2009).
194 Mordwinkin, N. M., Lee, A. S. & Wu, J. C. Patient-specific stem cells and
cardiovascular drug discovery. Jama 310, 2039-2040,
doi:10.1001/jama.2013.282409 (2013).
195 Walker, B. D. et al. Congenital and acquired long QT syndromes. The Canadian
journal of cardiology 19, 76-87 (2003).
196 Chevalier, P. et al. Non-invasive testing of acquired long QT syndrome:
evidence for multiple arrhythmogenic substrates. Cardiovascular research 50,
386-398 (2001).
197 Liang, P. et al. Drug screening using a library of human induced pluripotent
stem cell-derived cardiomyocytes reveals disease-specific patterns of
cardiotoxicity. Circulation 127, 1677-1691,
doi:10.1161/CIRCULATIONAHA.113.001883 (2013).
198 Wang, Y. et al. Genome editing of isogenic human induced pluripotent stem
cells recapitulates long QT phenotype for drug testing. Journal of the
American College of Cardiology 64, 451-459, doi:10.1016/j.jacc.2014.04.057
(2014).
199 Lafuente-Lafuente, C., Longas-Tejero, M. A., Bergmann, J. F. & Belmin, J.
Antiarrhythmics for maintaining sinus rhythm after cardioversion of atrial
fibrillation. The Cochrane database of systematic reviews 5, CD005049,
doi:10.1002/14651858.CD005049.pub3 (2012).
200 Brito-Martins, M., Harding, S. E. & Ali, N. N. beta(1)- and beta(2)-
adrenoceptor responses in cardiomyocytes derived from human embryonic
stem cells: comparison with failing and non-failing adult human heart. British
journal of pharmacology 153, 751-759, doi:10.1038/sj.bjp.0707619 (2008).
114
201 Lu, H. R., Marien, R., Saels, A. & De Clerck, F. Species plays an important role
in drug-induced prolongation of action potential duration and early
afterdepolarizations in isolated Purkinje fibers. Journal of cardiovascular
electrophysiology 12, 93-102 (2001).
202 Lian, Q., Chow, Y., Esteban, M. A., Pei, D. & Tse, H. F. Future perspective of
induced pluripotent stem cells for diagnosis, drug screening and treatment of
human diseases. Thrombosis and haemostasis 104, 39-44, doi:10.1160/TH10-
05-0269 (2010).
203 Kennedy, M., D'Souza, S. L., Lynch-Kattman, M., Schwantz, S. & Keller, G.
Development of the hemangioblast defines the onset of hematopoiesis in
human ES cell differentiation cultures. Blood 109, 2679-2687,
doi:10.1182/blood-2006-09-047704 (2007).
204 Hamill, O. P., Marty, A., Neher, E., Sakmann, B. & Sigworth, F. J. Improved
patch-clamp techniques for high-resolution current recording from cells and
cell-free membrane patches. Pflugers Archiv : European journal of physiology
391, 85-100 (1981).
205 Barry, P. H. & Lynch, J. W. Liquid junction potentials and small cell effects in
patch-clamp analysis. The Journal of membrane biology 121, 101-117 (1991).
206 Fluhler, E., Burnham, V. G. & Loew, L. M. Spectra, membrane binding, and
potentiometric responses of new charge shift probes. Biochemistry 24, 5749-
5755 (1985).
207 Fedorov, V. V. et al. Application of blebbistatin as an excitation-contraction
uncoupler for electrophysiologic study of rat and rabbit hearts. Heart rhythm
: the official journal of the Heart Rhythm Society 4, 619-626,
doi:10.1016/j.hrthm.2006.12.047 (2007).
208 Elliott, D. A. et al. NKX2-5(eGFP/w) hESCs for isolation of human cardiac
progenitors and cardiomyocytes. Nature methods 8, 1037-1040,
doi:10.1038/nmeth.1740 (2011).
209 Schwartz, R. J. & Olson, E. N. Building the heart piece by piece: modularity of
cis-elements regulating Nkx2-5 transcription. Development 126, 4187-4192
(1999).
210 Gassanov, N. et al. Retinoid acid-induced effects on atrial and pacemaker cell
differentiation and expression of cardiac ion channels. Differentiation;
research in biological diversity 76, 971-980, doi:10.1111/j.1432-
0436.2008.00283.x (2008).
211 Storms, R. W. et al. Isolation of primitive human hematopoietic progenitors
on the basis of aldehyde dehydrogenase activity. Proceedings of the National
Academy of Sciences of the United States of America 96, 9118-9123 (1999).
212 Evseenko, D. et al. Mapping the first stages of mesoderm commitment during
differentiation of human embryonic stem cells. Proceedings of the National
Academy of Sciences of the United States of America 107, 13742-13747,
doi:10.1073/pnas.1002077107 (2010).
213 Prall, O. W. et al. An Nkx2-5/Bmp2/Smad1 negative feedback loop controls
heart progenitor specification and proliferation. Cell 128, 947-959,
doi:10.1016/j.cell.2007.01.042 (2007).
115
214 Dubois, N. C. et al. SIRPA is a specific cell-surface marker for isolating
cardiomyocytes derived from human pluripotent stem cells. Nature
biotechnology 29, 1011-1018, doi:10.1038/nbt.2005 (2011).
215 Xie, Y., Sato, D., Garfinkel, A., Qu, Z. & Weiss, J. N. So little source, so much sink:
requirements for afterdepolarizations to propagate in tissue. Biophysical
journal 99, 1408-1415, doi:10.1016/j.bpj.2010.06.042 (2010).
216 Cabo, C. et al. Vortex shedding as a precursor of turbulent electrical activity in
cardiac muscle. Biophysical journal 70, 1105-1111, doi:10.1016/S0006-
3495(96)79691-1 (1996).
217 Bray, M. A. & Wikswo, J. P. Considerations in phase plane analysis for
nonstationary reentrant cardiac behavior. Physical review. E, Statistical,
nonlinear, and soft matter physics 65, 051902 (2002).
218 Mansour, M. et al. Left-to-right gradient of atrial frequencies during acute
atrial fibrillation in the isolated sheep heart. Circulation 103, 2631-2636
(2001).
219 Gonzalez, H., Nagai, Y., Bub, G., Glass, L. & Shrier, A. Reentrant waves in a ring
of embryonic chick ventricular cells imaged with a Ca2+ sensitive dye. Bio
Systems 71, 71-80 (2003).
220 Iravanian, S. et al. Functional reentry in cultured monolayers of neonatal rat
cardiac cells. American journal of physiology. Heart and circulatory physiology
285, H449-456, doi:10.1152/ajpheart.00896.2002 (2003).
221 Gibson, J. K., Yue, Y., Bronson, J., Palmer, C. & Numann, R. Human stem cell-
derived cardiomyocytes detect drug-mediated changes in action potentials
and ion currents. Journal of pharmacological and toxicological methods 70,
255-267, doi:10.1016/j.vascn.2014.09.005 (2014).
222 Hewlett, A. W. Some Clinical Aspects of Auricular Fibrillation: A Critical
Review. California and western medicine 22, 479-483 (1924).
223 Janse, M. J. & Rosen, M. R. History of arrhythmias. Handbook of experimental
pharmacology, 1-39 (2006).
224 Veerman, C. C. et al. Immaturity of Human Stem-Cell-Derived Cardiomyocytes
in Culture: Fatal Flaw or Soluble Problem? Stem cells and development,
doi:10.1089/scd.2014.0533 (2015).
225 Magyar, J. et al. Effects of endothelin-1 on calcium and potassium currents in
undiseased human ventricular myocytes. Pflugers Archiv : European journal
of physiology 441, 144-149 (2000).
226 Devalla, H. D. et al. Atrial-like cardiomyocytes from human pluripotent stem
cells are a robust preclinical model for assessing atrial-selective
pharmacology. EMBO molecular medicine 7, 394-410,
doi:10.15252/emmm.201404757 (2015).
227 Spach, M. S., Dolber, P. C. & Anderson, P. A. Multiple regional differences in
cellular properties that regulate repolarization and contraction in the right
atrium of adult and newborn dogs. Circulation research 65, 1594-1611
(1989).
228 Nygren, A., Lomax, A. E. & Giles, W. R. Heterogeneity of action potential
durations in isolated mouse left and right atria recorded using voltage-
116
sensitive dye mapping. American journal of physiology. Heart and circulatory
physiology 287, H2634-2643, doi:10.1152/ajpheart.00380.2004 (2004).
229 Feng, J., Yue, L., Wang, Z. & Nattel, S. Ionic mechanisms of regional action
potential heterogeneity in the canine right atrium. Circulation research 83,
541-551 (1998).
230 Li, D., Zhang, L., Kneller, J. & Nattel, S. Potential ionic mechanism for
repolarization differences between canine right and left atrium. Circulation
research 88, 1168-1175 (2001).
231 Ridler, M. E., Lee, M., McQueen, D., Peskin, C. & Vigmond, E. Arrhythmogenic
consequences of action potential duration gradients in the atria. The
Canadian journal of cardiology 27, 112-119, doi:10.1016/j.cjca.2010.12.002
(2011).
232 Jalife, J., Berenfeld, O. & Mansour, M. Mother rotors and fibrillatory
conduction: a mechanism of atrial fibrillation. Cardiovascular research 54,
204-216 (2002).
233 Jansen, J. A., van Veen, T. A., de Bakker, J. M. & van Rijen, H. V. Cardiac
connexins and impulse propagation. Journal of molecular and cellular
cardiology 48, 76-82, doi:10.1016/j.yjmcc.2009.08.018 (2010).
234 Vreeker, A. et al. Assembly of the cardiac intercalated disk during pre- and
postnatal development of the human heart. PloS one 9, e94722,
doi:10.1371/journal.pone.0094722 (2014).
235 Kadota, S. et al. Development of a reentrant arrhythmia model in human
pluripotent stem cell-derived cardiac cell sheets. European heart journal 34,
1147-1156, doi:10.1093/eurheartj/ehs418 (2013).
236 Wijffels, M. C., Dorland, R., Mast, F. & Allessie, M. A. Widening of the excitable
gap during pharmacological cardioversion of atrial fibrillation in the goat:
effects of cibenzoline, hydroquinidine, flecainide, and d-sotalol. Circulation
102, 260-267 (2000).
237 Hiraoka, M., Sunami, A., Fan, Z. & Sawanobori, T. Multiple ionic mechanisms
of early afterdepolarizations in isolated ventricular myocytes from guinea-
pig hearts. Annals of the New York Academy of Sciences 644, 33-47 (1992).
238 Yamamoto, W. et al. Effects of the selective KACh channel blocker NTC-801
on atrial fibrillation in a canine model of atrial tachypacing: comparison with
class Ic and III drugs. Journal of cardiovascular pharmacology 63, 421-427,
doi:10.1097/FJC.0000000000000065 (2014).
239 Herrera, D. et al. A single residue in the S6 transmembrane domain governs
the differential flecainide sensitivity of voltage-gated potassium channels.
Molecular pharmacology 68, 305-316, doi:10.1124/mol.104.009506 (2005).
240 Ramos, E. & O'Leary M, E. State-dependent trapping of flecainide in the
cardiac sodium channel. The Journal of physiology 560, 37-49,
doi:10.1113/jphysiol.2004.065003 (2004).
241 Nattel, S., Burstein, B. & Dobrev, D. Atrial remodeling and atrial fibrillation:
mechanisms and implications. Circulation. Arrhythmia and electrophysiology
1, 62-73, doi:10.1161/CIRCEP.107.754564 (2008).
117
242 Fujiki, A. & Inoue, H. Pharmacological cardioversion of long-lasting atrial
fibrillation. Circulation journal : official journal of the Japanese Circulation
Society 71 Suppl A, A69-74 (2007).
243 Yanagi, K. et al. Hyperpolarization-activated cyclic nucleotide-gated channels
and T-type calcium channels confer automaticity of embryonic stem cell-
derived cardiomyocytes. Stem cells 25, 2712-2719,
doi:10.1634/stemcells.2006-0388 (2007).
244 Ivashchenko, C. Y. et al. Human-induced pluripotent stem cell-derived
cardiomyocytes exhibit temporal changes in phenotype. American journal of
physiology. Heart and circulatory physiology 305, H913-922,
doi:10.1152/ajpheart.00819.2012 (2013).
245 Conde, D. & Baranchuk, A. Vernakalant for the conversion of atrial
fibrillation: the new kid on the block? Annals of noninvasive electrocardiology
: the official journal of the International Society for Holter and Noninvasive
Electrocardiology, Inc 19, 299-302, doi:10.1111/anec.12164 (2014).
246 Guerra, F., Matassini, M. V., Scappini, L., Urbinati, A. & Capucci, A. Intravenous
vernakalant for the rapid conversion of recent onset atrial fibrillation:
systematic review and meta-analysis. Expert review of cardiovascular therapy
12, 1067-1075, doi:10.1586/14779072.2014.943662 (2014).
247 Savelieva, I., Graydon, R. & Camm, A. J. Pharmacological cardioversion of
atrial fibrillation with vernakalant: evidence in support of the ESC Guidelines.
Europace : European pacing, arrhythmias, and cardiac electrophysiology :
journal of the working groups on cardiac pacing, arrhythmias, and cardiac
cellular electrophysiology of the European Society of Cardiology 16, 162-173,
doi:10.1093/europace/eut274 (2014).
248 Wilhelms, M. et al. Benchmarking electrophysiological models of human
atrial myocytes. Frontiers in physiology 3, 487,
doi:10.3389/fphys.2012.00487 (2012).
249 Leyton-Mange, J. S. et al. Rapid cellular phenotyping of human pluripotent
stem cell-derived cardiomyocytes using a genetically encoded fluorescent
voltage sensor. Stem cell reports 2, 163-170,
doi:10.1016/j.stemcr.2014.01.003 (2014).
250 Cao, G. et al. Genetically targeted optical electrophysiology in intact neural
circuits. Cell 154, 904-913, doi:10.1016/j.cell.2013.07.027 (2013).
251 Haley, C. L. et al. Validation of a novel algorithm for quantification of the
percentage of signal fractionation in atrial fibrillation. Europace : European
pacing, arrhythmias, and cardiac electrophysiology : journal of the working
groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology
of the European Society of Cardiology 15, 447-452,
doi:10.1093/europace/eus361 (2013).
252 Lau, D. H. et al. Stability of complex fractionated atrial electrograms: a
systematic review. Journal of cardiovascular electrophysiology 23, 980-987,
doi:10.1111/j.1540-8167.2012.02335.x (2012).
253 Berenfeld, O. & Jalife, J. Complex fractionated atrial electrograms: is this the
beast to tame in atrial fibrillation? Circulation. Arrhythmia and
electrophysiology 4, 426-428, doi:10.1161/CIRCEP.111.964841 (2011).
118
254 Van Gelder, I. C. et al. Chronic atrial fibrillation. Success of serial
cardioversion therapy and safety of oral anticoagulation. Archives of internal
medicine 156, 2585-2592 (1996).
255 Corradi, D. Atrial fibrillation from the pathologist's perspective.
Cardiovascular pathology : the official journal of the Society for Cardiovascular
Pathology 23, 71-84, doi:10.1016/j.carpath.2013.12.001 (2014).
256 Pinho-Gomes, A. C., Reilly, S., Brandes, R. P. & Casadei, B. Targeting
inflammation and oxidative stress in atrial fibrillation: role of 3-hydroxy-3-
methylglutaryl-coenzyme a reductase inhibition with statins. Antioxidants &
redox signaling 20, 1268-1285, doi:10.1089/ars.2013.5542 (2014).
257 Jalife, J. & Kaur, K. Atrial remodeling, fibrosis, and atrial fibrillation. Trends in
cardiovascular medicine, doi:10.1016/j.tcm.2014.12.015 (2014).
258 Atienza, F., Martins, R. P. & Jalife, J. Translational research in atrial
fibrillation: a quest for mechanistically based diagnosis and therapy.
Circulation. Arrhythmia and electrophysiology 5, 1207-1215,
doi:10.1161/CIRCEP.111.970335 (2012).
259 Berenfeld, O. & Jalife, J. Mechanisms of atrial fibrillation: rotors, ionic
determinants, and excitation frequency. Cardiology clinics 32, 495-506,
doi:10.1016/j.ccl.2014.07.001 (2014).
260 Baher, A. et al. Short-term cardiac memory and mother rotor fibrillation.
American journal of physiology. Heart and circulatory physiology 292, H180-
189, doi:10.1152/ajpheart.00944.2005 (2007).
261 Grandi, E. et al. Human atrial action potential and Ca2+ model: sinus rhythm
and chronic atrial fibrillation. Circulation research 109, 1055-1066,
doi:10.1161/CIRCRESAHA.111.253955 (2011).
262 Qi, X. Y. et al. Fibroblast inward-rectifier potassium current upregulation in
profibrillatory atrial remodeling. Circulation research 116, 836-845,
doi:10.1161/CIRCRESAHA.116.305326 (2015).
263 Chen, P. S., Chen, L. S., Fishbein, M. C., Lin, S. F. & Nattel, S. Role of the
autonomic nervous system in atrial fibrillation: pathophysiology and therapy.
Circulation research 114, 1500-1515, doi:10.1161/CIRCRESAHA.114.303772
(2014).
264 Voigt, N., Makary, S., Nattel, S. & Dobrev, D. Voltage-clamp-based methods for
the detection of constitutively active acetylcholine-gated I(K,ACh) channels in
the diseased heart. Methods in enzymology 484, 653-675, doi:10.1016/B978-
0-12-381298-8.00032-0 (2010).
265 Shiroshita-Takeshita, A., Brundel, B. J., Lavoie, J. & Nattel, S. Prednisone
prevents atrial fibrillation promotion by atrial tachycardia remodeling in
dogs. Cardiovascular research 69, 865-875 (2006).
266 Cagli, K. E. et al. Plasma levels of tumor necrosis factor-alpha and its
receptors in patients with mitral stenosis and sinus rhythm undergoing
percutaneous balloon valvuloplasty. Heart Vessels 25, 131-137,
doi:10.1007/s00380-009-1175-9 (2010).
267 Li, J. et al. Role of inflammation and oxidative stress in atrial fibrillation.
Heart rhythm : the official journal of the Heart Rhythm Society 7, 438-444,
doi:S1547-5271(09)01380-0 [pii]
119
10.1016/j.hrthm.2009.12.009 (2010).
268 Qi, J. et al. Effects of potassium channel blockers on changes in refractoriness
of atrial cardiomyocytes induced by stretch. Experimental biology and
medicine 234, 779-784, doi:10.3181/0902-RM-45 (2009).
269 Kalifa, J. et al. Anti-arrhythmic effects of I (Na), I (Kr), and combined I (Kr)-I
(CaL) blockade in an experimental model of acute stretch-related atrial
fibrillation. Cardiovascular drugs and therapy / sponsored by the International
Society of Cardiovascular Pharmacotherapy 21, 47-53, doi:10.1007/s10557-
007-6001-y (2007).
270 Dobrev, D., Carlsson, L. & Nattel, S. Novel molecular targets for atrial
fibrillation therapy. Nature reviews. Drug discovery 11, 275-291,
doi:10.1038/nrd3682 (2012).
271 Ueda, N., Yamamoto, M., Honjo, H., Kodama, I. & Kamiya, K. The role of gap
junctions in stretch-induced atrial fibrillation. Cardiovascular research 104,
364-370, doi:10.1093/cvr/cvu202 (2014).
272 Boldt, A. et al. Fibrosis in left atrial tissue of patients with atrial fibrillation
with and without underlying mitral valve disease. Heart 90, 400-405 (2004).
273 Frustaci, A. et al. Histological substrate of atrial biopsies in patients with lone
atrial fibrillation. Circulation 96, 1180-1184 (1997).
274 Kostin, S. et al. Structural correlate of atrial fibrillation in human patients.
Cardiovascular research 54, 361-379 (2002).
275 Woods, C. E. & Olgin, J. Atrial fibrillation therapy now and in the future: drugs,
biologicals, and ablation. Circulation research 114, 1532-1546,
doi:10.1161/CIRCRESAHA.114.302362 (2014).
276 Witty, A. D. et al. Generation of the epicardial lineage from human pluripotent
stem cells. Nature biotechnology 32, 1026-1035, doi:10.1038/nbt.3002
(2014).
277 Thavandiran, N. et al. Design and formulation of functional pluripotent stem
cell-derived cardiac microtissues. Proceedings of the National Academy of
Sciences of the United States of America 110, E4698-4707,
doi:10.1073/pnas.1311120110 (2013).
278 Qi, X. Y. et al. Fibroblast Inward-Rectifier Potassium Current Upregulation in
Profibrillatory Atrial Remodeling. Circulation research,
doi:10.1161/CIRCRESAHA.116.305326 (2015).
279 Suarez, G. & Meyerrose, G. Heart failure and galectin 3. Annals of translational
medicine 2, 86, doi:10.3978/j.issn.2305-5839.2014.09.10 (2014).
280 Pellman, J., Lyon, R. C. & Sheikh, F. Extracellular matrix remodeling in atrial
fibrosis: mechanisms and implications in atrial fibrillation. Journal of
molecular and cellular cardiology 48, 461-467,
doi:10.1016/j.yjmcc.2009.09.001 (2010).
281 Lin, C. S. & Pan, C. H. Regulatory mechanisms of atrial fibrotic remodeling in
atrial fibrillation. Cell Mol Life Sci 65, 1489-1508 (2008).
282 de Bakker, J. M. et al. Slow conduction in the infarcted human heart. 'Zigzag'
course of activation. Circulation 88, 915-926 (1993).
283 Hanna, N., Cardin, S., Leung, T. K. & Nattel, S. Differences in atrial versus
ventricular remodeling in dogs with ventricular tachypacing-induced
120
congestive heart failure. Cardiovascular research 63, 236-244,
doi:10.1016/j.cardiores.2004.03.026 (2004).
284 Nakajima, H. et al. Atrial but not ventricular fibrosis in mice expressing a
mutant transforming growth factor-beta(1) transgene in the heart.
Circulation research 86, 571-579 (2000).
285 Tucker, N. R. & Ellinor, P. T. Emerging directions in the genetics of atrial
fibrillation. Circulation research 114, 1469-1482,
doi:10.1161/CIRCRESAHA.114.302225 (2014).
286 Gonzalez, F. et al. An iCRISPR platform for rapid, multiplexable, and inducible
genome editing in human pluripotent stem cells. Cell stem cell 15, 215-226,
doi:10.1016/j.stem.2014.05.018 (2014).
287 Yang, Y. Q. et al. Prevalence and spectrum of GATA6 mutations associated
with familial atrial fibrillation. International journal of cardiology 155, 494-
496, doi:10.1016/j.ijcard.2011.12.091 (2012).
288 Zhao, R. et al. Loss of both GATA4 and GATA6 blocks cardiac myocyte
differentiation and results in acardia in mice. Developmental biology 317,
614-619, doi:10.1016/j.ydbio.2008.03.013 (2008).
289 Remond, M. C. et al. GATA6 reporter gene reveals myocardial phenotypic
heterogeneity that is related to variations in gap junction coupling. American
journal of physiology. Heart and circulatory physiology 301, H1952-1964,
doi:10.1152/ajpheart.00635.2011 (2011).
290 Arnolds, D. E. et al. TBX5 drives Scn5a expression to regulate cardiac
conduction system function. The Journal of clinical investigation 122, 2509-
2518, doi:10.1172/JCI62617 (2012).
291 Postma, A. V. et al. A gain-of-function TBX5 mutation is associated with
atypical Holt-Oram syndrome and paroxysmal atrial fibrillation. Circulation
research 102, 1433-1442, doi:10.1161/CIRCRESAHA.107.168294 (2008).
292 Zang, X. et al. SNP rs3825214 in TBX5 is associated with lone atrial
fibrillation in Chinese Han population. PloS one 8, e64966,
doi:10.1371/journal.pone.0064966 (2013).
293 Sinner, M. F. et al. Integrating genetic, transcriptional, and functional analyses
to identify 5 novel genes for atrial fibrillation. Circulation 130, 1225-1235,
doi:10.1161/CIRCULATIONAHA.114.009892 (2014).
294 Sartiani, L. et al. Developmental changes in cardiomyocytes differentiated
from human embryonic stem cells: a molecular and electrophysiological
approach. Stem cells 25, 1136-1144, doi:10.1634/stemcells.2006-0466
(2007).
295 Nunes, S. S. et al. Biowire: a platform for maturation of human pluripotent
stem cell-derived cardiomyocytes. Nature methods 10, 781-787,
doi:10.1038/nmeth.2524 (2013).
296 Zhang, M. et al. Universal Cardiac Induction of Human Pluripotent Stem Cells
in 2D and 3D formats - Implications for In-Vitro Maturation. Stem cells,
doi:10.1002/stem.1964 (2015).
297 Yang, X. et al. Tri-iodo-l-thyronine promotes the maturation of human
cardiomyocytes-derived from induced pluripotent stem cells. Journal of
121
molecular and cellular cardiology 72, 296-304,
doi:10.1016/j.yjmcc.2014.04.005 (2014).
122
Appendix
Appendix Figure 1: Intracellular recordings of a “ventricular” cell sheet performed with a
high resistance (~30 MΩ) thin walled glass micropipette. Sharp spikes represent pacing
artifact. The cell sheet was paced from the edge of one side of the cell sheet, and a cell
on the opposite side was impaled for this recording. The conduction velocity was
calculated to be ~4 cm/s for this recording. The resting membrane potential, upstroke
velocity and action potential amplitude mimic those recorded from single cell patch
clamping studies. In particular, the resting membrane potential sits at ~-55 mV,
essentially the same as seen in single cell studies.
123
Appendix Figure 2: Activation maps generated from recordings of an “atrial” cell sheet
and the sequential uptitration of flecainide dosing using DI-4ANEPPS and an EMCCD.
Activation maps demonstrate the dose dependent effects of flecainide on the cycle length
and the number of excitation wavelets in the cell sheet. Initially a stable rotor is induced
on the left. After the addition of 5 μM flecainide, the cycle length is substantially
reduced and the curvature of the rotor is decreased without affecting the location of the
phase singularity (PS)(labeled as a white star on figure). After the addition of 10 μM
flecainide, two distinct rotors appear, each rotating around their own PS independently
and causing wave collision along sites of wavefront interaction. Wave collision did not
generate more wavelets in this cell sheet, and instead the activation wavefront
extinguished at the site of collision.
124
Appendix Figure 3: Optical Action Potentials (OAPs) recorded using the Arclight hiPSC cell line differentiated into cardiomyocytes and recorded on an EMCCD. Cells contain a genetically encoded voltage sensor attached to eGFP which reports a change in membrane voltage as changes in fluorescence. Note the 20% change in fluorescent signal intensity (Y axis) demarcating depolarization of the hiPSC derived cardiomyocytes. Note also the absence of photobleaching over 20 seconds of recording time, overcoming a major limitation of the voltage sensitive dyes. This was repeated on many occasions on the same day and on sequential days for up to 1 week with preservation of signal.